CA2227815C - Apparatus and method for automated visual inspection of objects - Google Patents
Apparatus and method for automated visual inspection of objects Download PDFInfo
- Publication number
- CA2227815C CA2227815C CA002227815A CA2227815A CA2227815C CA 2227815 C CA2227815 C CA 2227815C CA 002227815 A CA002227815 A CA 002227815A CA 2227815 A CA2227815 A CA 2227815A CA 2227815 C CA2227815 C CA 2227815C
- Authority
- CA
- Canada
- Prior art keywords
- inspection
- target
- objects
- optic sensor
- targets
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Lifetime
Links
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
Abstract
The present invention relates to an apparatus and method for the automated inspection of objects for the presence of visual anomalies. The apparatus includes a first wide angle camera for acquiring an image of the object, a processor for detecting targets and generating target location from the image of the object, a two axis post-objective scanner associated with a high resolution camera for directing target image to the high resolution camera. The scanner provides a mechanism for saccadic emulation in combination with a high resolution video camera, which can be synchronized to the frame rate of the video camera for rapid high resolution examination of an object. The method according to the present invention comprises the steps of: capturing an image of at least a portion of an object; identifying visual anomalies as targets and detecting any targets on the object; generating location data for a detected target; emulating saccadic niovement to selectively capture a high resolution image of a target; and generating an output for analysis. The apparatus and method are particularly suited for the automated detection of defects, in particular parasites, in fish fillets. It is an advantage of the present invention that post-objective scanning permits the camera to remain stationary, while only the mirrors are moved, to acquire high resolution data of substantially only the targets, resulting in increased speed and reduced processing time.
Description
APPARATUS AND METHOD FOR AUTOMA'I'ED
VISUAL INSPECTION OF OB-JECTS
FTELD OF THE INVENTION
The present invention relates to a method and apparatus suitable for use in a niachine vision automated inspection station. In particular the present invention relates to a method and apparatus for rapidly detecting and inspecting visual anomalies in objects.
EIACKGROLfND OF THE INTVENTION
Machine vision has successfully been integrated into many manufacturing systems for automateci inspection and automated control. Video cameras capture images with adequate resolution, typically 512 x 512. The image data can be stored and processed by a computer. With a narrow field of view, the resolution is sufficient for examining small details or detecting flaws for quality control. However, inspection of large areas at high resolution requires handling a very large volume of data. The penalty for detailed automated inspection is the time necessary for image acquisition and for analysis of the data. Many automated inspection applications have not been economically feasible due to time constraints and computing costs.
Automated inspection at high resolution is quite common. However, systems proposed are too slow to examine large areas or numerous sites effectively.
For example, iJnited States Patent No. 5,517,235 issued to Wasserman, uses video cameras with zoom lenses to vary the field of view for inspecting printed circuit boards. A
smaller field of view, and herice higher resolution, is needed to examine boards or parts of boards with greater density of components. The inspection head is advanced across a viewing table to acquire a complete image. Multiple passes may be made to increase the resolution for examining critical areas flagged in the initial inspection. Physical movement of the head back and forth, and zoom lens control are relatively slow and not selective for examining specific locations. Even if only a small field of high resolution data is to be acquired, the head must pass over the complete object to locate the area.
VISUAL INSPECTION OF OB-JECTS
FTELD OF THE INVENTION
The present invention relates to a method and apparatus suitable for use in a niachine vision automated inspection station. In particular the present invention relates to a method and apparatus for rapidly detecting and inspecting visual anomalies in objects.
EIACKGROLfND OF THE INTVENTION
Machine vision has successfully been integrated into many manufacturing systems for automateci inspection and automated control. Video cameras capture images with adequate resolution, typically 512 x 512. The image data can be stored and processed by a computer. With a narrow field of view, the resolution is sufficient for examining small details or detecting flaws for quality control. However, inspection of large areas at high resolution requires handling a very large volume of data. The penalty for detailed automated inspection is the time necessary for image acquisition and for analysis of the data. Many automated inspection applications have not been economically feasible due to time constraints and computing costs.
Automated inspection at high resolution is quite common. However, systems proposed are too slow to examine large areas or numerous sites effectively.
For example, iJnited States Patent No. 5,517,235 issued to Wasserman, uses video cameras with zoom lenses to vary the field of view for inspecting printed circuit boards. A
smaller field of view, and herice higher resolution, is needed to examine boards or parts of boards with greater density of components. The inspection head is advanced across a viewing table to acquire a complete image. Multiple passes may be made to increase the resolution for examining critical areas flagged in the initial inspection. Physical movement of the head back and forth, and zoom lens control are relatively slow and not selective for examining specific locations. Even if only a small field of high resolution data is to be acquired, the head must pass over the complete object to locate the area.
Closer examination of specific sites of interest is also proposed as in United States Patent No. 5,051,825 issued to Cochran and Austin, which uses a first video camera to determine the position and orientation of the article using geometrical determination, and a second video camera which simultaneously captures an image of higher resolution while the article is advanced across the field of view. This is a slow process.
Acquiring a full iinage at high resolution presents greater time and system capability demands.
United States Patent No. 4,872,052 issued to Liudzius, et al., discloses a system for examining seini-conductors which assembles images from multiple angles of view for examining a three-dimensional object at low resolution. The system also includes a high resolution carnera which receives location information from the low resolution image to adjust the im,age registration to the standard comparison data. The high resolution camera is moveable on an X-Y plane parallel to that of the semi-conductor in order to acquire target image data. It is provided with a focus compensating device for fme adjustment to the high magnification without changing the focal length. Moving the camera across the object and further manipulating a focus compensating device are both complicated inechanically and slow. It is not anticipated by this prior art design to examine numerous targets at high resolution.
Slow nlechanical positioning or acquisition of large volumes of unnecessary data have rendered prior art automated inspection systems too slow to be appropriate for high resolution examination of numerous sites, or for relatively high speed automated processing.
Scanners for precise and rapid positioning of optic mirrors have been used for raster scanning to acquire detailed composite information, such as in confocal n:iicroscopes, CAT SCAN and MRI medical imaging equipment, and in industry for example in three dimensional imaging techniques. The raster scan makes a complete pass across the object assembling data line by line. A more complex scanner system includes two axes of rotation, offering precise positioning to any point across a plane vrithin angular limitations. Dual axis scanners have been used for directing lasers, for instance in targeting systems, and light show demonstrations. A dual axis scanner generally combines a high quality optic mirror mounted on a magnetically driven armature withi a position detector and a digital driver. The accuracy of dual axis scanners can be within a few microradians.
Use of a two axis scanner is proposed in United States Patent No. 5,608,563 issued to Matsumura et al. in order to rapidly position a scanner to lines which contain data only. The scanner is used in cooperation with a projection apparatus rather than for data acquisition. ']['ime is saved by not scanning lines which do not contain data.
However, the location of lines containing data is known in advance and stored in the system before scanning begins. The system cannot be used to locate previously unknown targets.
In the inspection of regular objects a comparison to a reference standard or template is easily made. Regular objects are well suited for geometrical analysis and standard connparison. Once position registration is established, target location information can be pre-programmed. The elements of interest have known configurations and positions, reducing processing time and scanning requirements.
Objects, such as food and textile products in particular, can be highly variable in s:hape and texture and pose a significantly more difficult problem for machine vision inspection. Ini some cases parameters can be prescribed, or statistical analysis can predict reliable results. Foreign objects, however, defy any classification of this kind. Foreign objects may have unpredictable size, shape, material, position, orientation, and number vvhich require a more adaptable approach to automated detection and analysis leading to identification. The more complex processing and resultant time required make rapid dletection and acquisition of targets image data more important for achieving an efficient system.
In both cases the rapid detection of targets, acquisition of data and reduced Firocessing tirne are important in producing an effective automated system.
Human visual inspection is commonly used for quality control for rapidly identifying superficial imperfections or blemishes. The eye is quickly drawn to irregularities from a broad scan of the object. The response is referred to as an attention IIiechanism. Features, such as motion and contrast, prunarily detected by human vision, a:re of interest to quantify in order to emulate the attention mechanism in automated systems. Other features such as color, texture and alignment can be used similarly by automated systems.
The present invention was developed for the inspection of fish fillets for defects, primarily parasites. This is an area which has been difficult to automate.
Most production functions in fin fish processing have been fully automated, but defect iiispection still requires skilled quality inspection personnel. Surface and near surface parasites are currently detected by human visual examination on a candling table which provides back: lighting through the fish fillet and highlights characteristic shadows that correspond to parasites. This is, obviously, a labor-intensive operation which adds to the production costs.
This application presents significant additional challenges to automated inspection.
T'he fillets do not have a regular shape or size. Objects of interest must be located and distinguished from other visual features such as blood spots, traces of skin or strands of sitomach lining which often have quite similar characteristics. Unlike other industrial inspection applications, the parasites do not have a fixed shape, size or number and are often transparent offering very low contrast. To make determinations of this kind, relatively hig.h resolution is required which can match human capabilities. To be commercially useful, analysis must occur in a relatively short time scale.
Human analysis nequires approximately 3 seconds per side of each fillet. An automated system with a throughput of one fillet per second is desirable for economic substitution.
The liunitation to providing high resolution imaging at human equivalent or even closer detail is the shear volume of data such an operation would create. A
high resolution image of a whole fillet would be too large for processing and analysis in a useful time scale. A fillet of an estimated 200 mm x 400 mm, at human equivalent resolution of 5 lines per mm, would generate an array of approximately 3000 x pixels. The sensors necessary to pick up such a large array are very expensive. A system capable of capturing and processing data at the necessary rate cannot be economically irnplemented. Emulation of human eye performance cannot be achieved using brute force.
To achieve human equivalent resolution, it has been proposed in the present irivention to mimic human visual processing. The human eye has a small central color sensitive area made up of cones capable of very high resolution. The majority of visual sensors which comprise the peripheral vision system are lower resolution monochrome rods. In human visual inspection, a broad scan by the peripheral vision system uses the attention mechanism to provide an initial identification of anomalies. The central vision sYstem is then used to examine anomalies in detail for final identification.
Eye niovement, known as saccadic fixation, jumps to relocate the eye position approximately 3-10 times per second having the effect of assembling a composite view of successive high resolution images. Imitating this two stage vision hierarchy involves collection of low resolution scene information over a wide area (typically 0.6 radian solid angle), providing an attention mechanism and directing a narrow field of view (about 17 mrads at a resolution of about 0.1 mrad) for close inspection of features of interest to assemble substantially complete high resolution information. Human eye saccadic fixation rate at about 10Hz has been emulated in a camera orienting system known as Agile Eye.
Flowever, the selective acquisition of high resolution information at relatively high speed (30-60 Hz) has not previously been achieved.
The present invention utilizing the post-objective scanning design to provide saccadic emulation, imitates human visual acquisition at an order of magnitude faster.
By achieving a 30-60 Hz acquisition rate, data acquisition can be matched to the normal fi-ame rate of a video camera, as commonly available, to capture 30 frames or 60 fields per second. Thus a relatively simple and inexpensive system is optimized to view an entirely different scene with each frame or field.
SUMMARY OF THE INVENTION
The present invention has found that significant savings in time can be achieved using the rapid and precise positioning capability of a two axis post-objective scanner in cooperation with a visual image sensor. Physical positioning time is reduced, and only selected high resolution data is acquired for processing. A video camera with a narrow field of view can be directed selectively to acquire only necessary target data, without acquiring high resolution image data of the complete object or scanning past the complete object thus reducing the total volume of data which must be handled. Preferably, the system according to the present invention is used in combination with an additional video camera having a wide field of view and an initial attention mechanism, such as blob analysis or other known image processing routines, to generate locations for closer inspection by the high resolution system.
Accordingly, the present invention provides an apparatus for the inspection of objects in an inspection field for the presence of a target, comprising:
a light source to illuminate the object;
first and second optic sensor means each being directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor means for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor; and second processor means associated with the second optic sensor means for generating an output of the target image data for analysis.
Desirably, the positioning means associated with the second optic sensor means comprises a post-objective scanner, the positioning means associated with the second optic sensor means comprises a post-objective scanner, the target includes an inspection field having a plane, the post-objective scanner is selectively directable in two axes toward a location on the plane in the inspection field, the post-objective scanner is directable sequentially towards each target location, the post-objective scanner has a random access of less than 33 milliseconds, and the post-objective scanner has a random access of about 3 milliseconds.
In a preferred embodiment, the present invention further includes a wide angle optic sensor means to acquire image data of the object to provide to the first processor means for detecting any target and for generating location data for each target.
In a preferred embodiment, the present invention further includes a high resolution optic sensor means having a smaller field of view than that of the wide angle sensor means, wherein the high resolution image includes a target substantially in the absence of unnecessary data.
Preferably, the wide angle sensor means and the high resolution sensor means are charge coupled device video cameras, the target comprises a visual anomaly, the first processor means includes a blob analysis routine for detecting visual anomalies, the wide angle optic sensor means and the high resolution optic sensor means comprise the same apparatus operating in different modes, and the second optic sensor means includes a field of view and the field of view of the high resolution optic sensor means is variable in response to the target size.
The present invention is preferably for the inspection of defects, including parasites in fish fillets.
In a further embodiment, the present invention provides an apparatus for the inspection of objects for the presence of targets including an optic sensor, the improvement comprising positioning means for saccadic emulation associated with the optic sensor for directing image data of a target to the optic sensor having a post-objective scanner.
It is further preferable, the target includes an inspection field having a plane and the post-objective scanner is selectively directable in two axes towards any location on the plane in the inspection field, the post-objective scanner has a random access of less than 33 milliseconds, the post-objective scanner has a random access of about 3 milliseconds, the optic sensor comprises a charge coupled device video camera having a frame rate and the post-objective scanner has a random access rate which can be synchronized with the frame rate of the camera, and further includes a processor means including a blob analysis routine.
In an additional embodiment, the present invention comprises an apparatus for the inspection of objects for visual targets, comprising:
a light source;
a first video imaging means for acquiring a visual image having a field of view including at least a portion of an object and for transmitting image data to a central processor;
a second video imaging means for acquiring a second visual image having a field of view smaller than the field of view of the first video imaging means and for transmitting image data to a central processor;
a central processor for receiving image data from the first and second video imaging means and for generating an output for analysis including:
a frame grabber for converting the image data of first and second video imaging means to digital data and isolating a single image;
a digital signal processor assembly programmed for morphological manipulation of the image data from the first video imaging means to detect individual visual anomalies as targets and for generating location coordinates for each target; and a post-objective scanner associated with the second video imaging means including:
a mirror directable in two axes for selectively directing image data from a specified target location to the second video imaging means; and a driver for accessing location coordinates and directing the mirror towards a target location.
Desirably, the video imaging means comprises charge coupled device video cameras, the cameras having a frame rate, and the post-objective scanner has a random access rate which can be synchronized with the frame rate of the camera.
In another embodiment of the present invention comprises an automated station for the detection of defects including parasites in fish fillets, including:
a viewing head comprising:
first image recording means for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target; and second processor means for generating an output from the target image data for analysis.
A preferred method, according to the present invention, for the inspection of objects for the presence of targets comprises:
detecting any targets on the object;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target; and generating an output for analysis.
Preferably, the method according to the present invention further includes synchronizing the saccadic movement emulation to sequentially capture a high resolution image of each target detected at the frame rate of a video imaging means, and the method includes a first step of treating an object with a marker prior to inspection to identify any targets and a detection means is provided for cooperation with the selected marker for detecting the presence of a target.
A further preferred method according to the present invention comprises a method for the inspection of fish fillets for the presence of defects including parasites, comprising the steps of:
9a capturing an image of at least a portion of a fillet;
detecting any defects or potential parasites as targets;
generating a location for a detected target;
emulating saccadic movement to selectively capture a higher resolution image of a target;
processing target image data to determine whether a target is a defect or parasite; and identifying a fillet having a defect or parasite.
In another embodiment of the present invention there is provided an apparatus for the inspection of objects in an inspection field for the presence of a target comprising:
a light source;
first and second optic sensor means each directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor means whereby the high resolution target image data is acquired and processed generally immediately following the generation of target location data by the first optic sensor and the first processor means; and second processor means associated with the second optic sensor for generating an output of the target image data for analysis.
Preferably, the positioning means associated with the optic sensor means comprises a post-objective scanner, the post-objective scanner is selectively directable in two axes toward a location on a plane in the inspection field, is directable sequentially towards each target location, and has a random access of less than 33 milliseconds.
It is further preferable the scanner has a random access of about 3 milliseconds.
9b In the above embodiment it is preferable the first optic sensor means is a wide angle optic sensor means provided to acquire image data of the object to provide to the first processor means for detecting any target and for generating location data for each target, the wide angle optic sensor means captures low resolution image data of the object, and the second optic sensor means comprises a high resolution optic sensor means having a smaller field of view than that of the wide angle sensor means, the wide angle sensor means and the high resolution sensor means are charge coupled device video cameras, and the targets comprise visual anomalies.
Moreover, the first processor means desirably includes a blob analysis routine for detecting the visual anomalies.
In addition, it is preferred in the above embodiment, the wide angle optic sensor means and the high resolution optic sensor means comprise the same apparatus operating in different modes, the high resolution optic sensor means includes a field of view and the field of view of the high resolution optic sensor means is variable in response to the target size, and the high resolution image includes a target substantially in the absence of unnecessary data.
In another embodiment of the present invention there is provided a method for the inspection of objects for the presence of targets comprising the steps of providing an apparatus as defined above;
detecting any targets on the object with the first optic sensor means;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target with the second sensor means generally immediately following the step of generating location data and without moving the object; and, generating an output for analysis.
Preferably, saccadic movement emulation step comprises a video imaging means, having a frame rate, and when there is a sequence of targets detected the capture of the high resolution image of each target detected is sequentially synchronized to the frame rate of a 9c video imaging means, and the method includes a first step of treating an object with a marker prior to inspection to identify any targets and a detection means is provided for cooperation with the selected marker for detecting the presence of a target, or the method includes a first step of capturing a first image of at least a portion of an object from which any targets can be detected.
Further, it is desirable the above embodiment includes identifying visual anomalies as targets for detection.
Moreover, it is desirable the object comprises a fish fillet, and the targets include parasites, the step of generating an output for analysis includes the step of processing target image data to determine whether the target is a defect or a parasite and identifying a fillet as having a defect or parasite.
In another embodiment of the present invention, there is provided an automated station for the detection of defects including parasites in fish fillets including:
a viewing head comprising:
first image recording means for viewing the object concurrently with the first image recording means and for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target;
9d second processor means for generating an output from the target image data for analysis; and wherein the first and second image recording means, and the first and second processor means operate generally one immediately following the other to emulate the functioning of a human eye to generally nearly simultaneously or in rapid succession identify targets within a broad field of view and resolve such targets within a narrow field of view at higher resolution.
It is an advantage of the present invention to provide the capability to detect targets, and to be able to selectively isolate targets for detailed examination. Speed is gained according to the present invention by acquiring only necessary high resolution data for processing and analysis. The system advantageously reduces the amount of data to be processed and consequently the time required. Post-objective scanning permits the camera to remain stationary as only the mirrors need to be moved. As a result, the speed is increased to accurately enable image acquisition synchronous with the frame rate of a standard video camera, and the driving power is reduced.
It is a further advantage to achieve fast and accurate performance utilizing commonly available, and hence cost effective, hardware.
The present invention is suitable for many visual examination tasks, particularly where multiple target acquisition is necessary, or in the examination of a relatively large field in a short time period. This system can be adapted for use in other examination tasks where identification in complex and irregular scenes must be made, such as other areas of food processing, inspection of natural materials, and medical diagnostics.
BRIEF DESCRIPTION OF THE FIGURES
The invention will be readily understood by the following description of a preferred embodiment with reference to the following drawings in which:
Figure 1 is a schematic illustration of a bench scale embodiment of the present invention;
Figure 2 is a cut-away view of the scan head of Figure 1;
Figure 3 is a schematic illustration of the embodiment of Figure 1 showing the logical process path;
Figure 4 is a schematic illustration of an embodiment of the present invention for the inspection of fish fillets; and Figure 5 is a flow chart schematically illustrating the first and second processor routines.
Like numerals are used throughout to designate like elements.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
An embodiment of the invention as a system is shown generally as 10 in Figure 1. As seen in a schematic bench scale model, a scan head 100 is adjustably mounted on a standard 101. The object 12 having test targets 14 is presented to the cameras 20, 30 (see Fig. 2) in the scan head 100, on table 16. The cameras 20, 30 are preferably video imaging means, such as CCD (charge coupled device) video cameras. Other equivalent image recording means may be substituted. A candling table 16 is provided for back-lighting a fillet 12 in the case of fish inspection. The image data acquired by the cameras can be displayed on the monitor 28, or in real time on the video monitor 29. General lighting 18 is also provided.
Lighting requirements, as known in the art, will change for each application. In this configuration the object 12 is presented manually and remains in the same position for image acquisition by the low and high resolution cameras 20, 30. The system 10 can also operate in a dynamic mode (seen in Figure 3) with the objects 12 travelling on an continuous motion conveyor 116 at a fixed rate.
As represented in Figure 3, the first camera 20 has a wide field of view Flwhich includes the entire object 12. A second camera 30 with a high magnification lens 31 has a narrow field of view FZ including only a reduced portion of the object 12 large enough to capture a potential target 14 at higher resolution substantially in the absence of unnecessary data. The narrow field of view F2 can be adjusted to accommodate the anticipated target size. If the speed of operation is reduced, the field of view FZ can be variable, for ivnstance using an automatic zoom lens, in response to the actual target size 14. Similarly, under reduced speed operation, a single camera can be adapted to work in low resolut;ion and high resolution modes using an automatic zoom lens to change the field of view.
A positioning device 32 comprises a high speed two-axis post-objective optical scanner for directing the field of view of the second camera 30. Seen in detail in Figure 2, the scan head 100 houses a first low resolution camera 20 and lens 21, and a high resolution carnera 30 which cooperates with a high magnification lens 31, an x-y scanner 32 having a niirror 34 on a mount 35 for two axis rotation.
Associated with the first CCD camera 20 for processing controlled by a CPU 26, seen schematically in Figure 3, is a frame grabber 22 for isolating a single digital image and a target locator 24 for detecting visual targets and for generating x-y coordinate target location information. Coordinate location information is provided to a scanner driver 36 for directing the field of view FZ of the second camera 30 towards a target. At 30-60Hz the scanner 32 directs new image data isolated by the frame grabber 22 at the firame rate of the camera 30. The CPU 26 also provides synchronization for the lighting 16, 18, cameiras 20, 30, and scanner 32. Processing operations associated with each camera are shown as separate loops, however the necessary hardware is actually common between theni.
Processing in the CPU 26 generates target location data, depending on the application, from a data table in combination with position registration information from a first sensor, a subtraction operation, location or image information from a first sensor.
The first sensor is provided to identify the presence of an object and the location of targets. Since the initial identification is not for visual analysis, other sensors, as available in the art, may be used to generate location information. Targets maybe pre-marked for detection with appropriate detection means, for instance by florescence in combination with a photodetector, or other well known markers. Advantageously target locations are icientified froin a first low resolution image acquired by a visual sensor, such as a line scan or CCD camera, for the inspection of irregular objects with a blob analysis routine, or comparable pattern recognition operation to detect targets and generate locations for detailed inspection. A frame grabber system 22 is required which can perform image digitization and image processing within one second. A programmable DSP
(digital signal processor) board or board with multiple DSPs 24, for blob analysis or other known irnage processing routines, may be interfaced with the frame grabber 22, for instance, to achieve the p:rocessing speed required to detect targets.
A second CCD camera 30 and high magnification lens 31 acquire isolated target irnage data, without unnecessary high resolution data. The scan system 32, 34, 36, allows the high resolution camera 30 to selectively obtain visual data from any location on the surface of the: fillet 12 without taking a high resolution image of the whole object 12.
The scanner 32 has a mirror 34 with two axis control at very rapid access time to direct light back to the high resolution camera 30. The scanner 32 and mirror 34 effectively position the fi"eld of view of the high resolution camera 30 anywhere in the object field.
In operation the first camera 20 acquires a low resolution image which is processed only once per object 12 to detect any targets 14. A frame grabber 22 isolates a single frame and converts the visual data to digital data. Processing in the central processor 26 using a blob analysis routine or equivalent target locator 24 then generates coordinates indicating the location of the targets 14. Coordinate data is used to drive the scanner 32, and mirror 34 sequentially to each location where the second camera 30 obtains high resolution image data of each target 14 within a narrow field of view. The h:igh resolution target image data can then be processed for further analysis, such as a ninning count, comparison, categorizing, acceptance or rejection, etc. The presence of known features may be subtracted leaving only foreign anomalies for inspection.
Visually targets are identified by shape, color, contrast, texture or other changes iri appearance which may be highlighted by lighting techniques such as lighting angle or back lighting, selected colored light, edge detection, misalignment, surface irregularity, color value, as, known in the art. Processing visual data to detect targets may be specified to certain pai-ameters or compared to a data library using algorithmic morphological rnanipulation such as blob analysis, thresholding, edge detection, dilation, erosion, subtraction, inter alia known in the art, which will vary depending on the specific application. 'The benefit in more specific target detection is that fewer target images niust be acquiired and processed, thus increasing throughput speed.
Figure 5 illustrates the first processor routine 300, and the second processor routine 400 in a fillet inspection application. In this preferred operation, as seen in 400, each target is inspected sequentially. A final output is made to accept or reject the fillet.
The system then returns to wait for the next fillet.
As seen in Figure 4 an embodiment including the present invention for fish irispection 200 includes a feed conveyor 216 adapted to stop at the inspection station or to advance the fillet 12 through at a controlled rate. Fillets 12 may be supplied from an additional conveyor upstream (not shown), or by manual placement. The inspection station comprises a candling light source 218, a scan head 100 (housed in a protective upper cabinet 219 shrouding the inspection field, and preferably separated by glass from the fish handling environment) and a computerized control interface 228. Once the inspection is complete the fillet 12 is advanced by the conveyor 216 and classified according to the results. This is represented by a mechanical deflector 220 and separate containers 222, 223. Separate outfeed conveyors (not shown) may alternatively be provided to convey the separated fillets for further processing according to the inspection results.
"ne scan head 100 for parasite detection uses two low resolution (eg. 512 x p:ixel) CCD viideo cameras, preferably having 24 bit color capability. For the present application, the lowest acceptable resolution to detect all potential targets in the low resolution camera 20 is 550 m or 0.5 lines per mm, for instance using a 6 mm lens. In this application a resolution of 78 m or 5 lines per mm is needed for reliable d:iscrimination. of features using a 60-300 mm lens 31 with the high resolution camera 30.
This provides a first field of view of approximately 200 x 400 mm and a narrow field of view of approximately 10-40 mm square. A scanner controllable over 20 at 3 msec.
access time is used in the present embodiment. The mirror 34 is 10mm, selected to cooperate with the aperture of the high resolution camera 30. Intense illumination is provided to compensate for a relatively small mirror which reduces drive power requirements and increases positioning speed. In the fillet inspection application approximately 20-30 targets per fillet must be inspected at high resolution in the one second throughput time. The elements used in this application have been selected as commercially available components for cost control. As apparent to a person sldlled in the art, equivalent components can be substituted for specific applications as needed.
The device according to the present invention has also been adapted for the icientification of visual blemishes in other applications, such as paint and other surface treatments in manufactured articles. Changes in color, texture, contrast and shape are easily identified. For large articles a series of inspections may be needed to inspect the complete object.
The present invention creates an output which can be used as a culling mechanism to reduce the number of objects for human inspection, as a sorting system by numerical tlireshold, as a search system for features in a visual data set, for off-site human irispection or with a pattern recognition and analysis system for further automated processing. Other applications of the invention as defined in the appended claims will be apparent to persons of skill in the art.
Acquiring a full iinage at high resolution presents greater time and system capability demands.
United States Patent No. 4,872,052 issued to Liudzius, et al., discloses a system for examining seini-conductors which assembles images from multiple angles of view for examining a three-dimensional object at low resolution. The system also includes a high resolution carnera which receives location information from the low resolution image to adjust the im,age registration to the standard comparison data. The high resolution camera is moveable on an X-Y plane parallel to that of the semi-conductor in order to acquire target image data. It is provided with a focus compensating device for fme adjustment to the high magnification without changing the focal length. Moving the camera across the object and further manipulating a focus compensating device are both complicated inechanically and slow. It is not anticipated by this prior art design to examine numerous targets at high resolution.
Slow nlechanical positioning or acquisition of large volumes of unnecessary data have rendered prior art automated inspection systems too slow to be appropriate for high resolution examination of numerous sites, or for relatively high speed automated processing.
Scanners for precise and rapid positioning of optic mirrors have been used for raster scanning to acquire detailed composite information, such as in confocal n:iicroscopes, CAT SCAN and MRI medical imaging equipment, and in industry for example in three dimensional imaging techniques. The raster scan makes a complete pass across the object assembling data line by line. A more complex scanner system includes two axes of rotation, offering precise positioning to any point across a plane vrithin angular limitations. Dual axis scanners have been used for directing lasers, for instance in targeting systems, and light show demonstrations. A dual axis scanner generally combines a high quality optic mirror mounted on a magnetically driven armature withi a position detector and a digital driver. The accuracy of dual axis scanners can be within a few microradians.
Use of a two axis scanner is proposed in United States Patent No. 5,608,563 issued to Matsumura et al. in order to rapidly position a scanner to lines which contain data only. The scanner is used in cooperation with a projection apparatus rather than for data acquisition. ']['ime is saved by not scanning lines which do not contain data.
However, the location of lines containing data is known in advance and stored in the system before scanning begins. The system cannot be used to locate previously unknown targets.
In the inspection of regular objects a comparison to a reference standard or template is easily made. Regular objects are well suited for geometrical analysis and standard connparison. Once position registration is established, target location information can be pre-programmed. The elements of interest have known configurations and positions, reducing processing time and scanning requirements.
Objects, such as food and textile products in particular, can be highly variable in s:hape and texture and pose a significantly more difficult problem for machine vision inspection. Ini some cases parameters can be prescribed, or statistical analysis can predict reliable results. Foreign objects, however, defy any classification of this kind. Foreign objects may have unpredictable size, shape, material, position, orientation, and number vvhich require a more adaptable approach to automated detection and analysis leading to identification. The more complex processing and resultant time required make rapid dletection and acquisition of targets image data more important for achieving an efficient system.
In both cases the rapid detection of targets, acquisition of data and reduced Firocessing tirne are important in producing an effective automated system.
Human visual inspection is commonly used for quality control for rapidly identifying superficial imperfections or blemishes. The eye is quickly drawn to irregularities from a broad scan of the object. The response is referred to as an attention IIiechanism. Features, such as motion and contrast, prunarily detected by human vision, a:re of interest to quantify in order to emulate the attention mechanism in automated systems. Other features such as color, texture and alignment can be used similarly by automated systems.
The present invention was developed for the inspection of fish fillets for defects, primarily parasites. This is an area which has been difficult to automate.
Most production functions in fin fish processing have been fully automated, but defect iiispection still requires skilled quality inspection personnel. Surface and near surface parasites are currently detected by human visual examination on a candling table which provides back: lighting through the fish fillet and highlights characteristic shadows that correspond to parasites. This is, obviously, a labor-intensive operation which adds to the production costs.
This application presents significant additional challenges to automated inspection.
T'he fillets do not have a regular shape or size. Objects of interest must be located and distinguished from other visual features such as blood spots, traces of skin or strands of sitomach lining which often have quite similar characteristics. Unlike other industrial inspection applications, the parasites do not have a fixed shape, size or number and are often transparent offering very low contrast. To make determinations of this kind, relatively hig.h resolution is required which can match human capabilities. To be commercially useful, analysis must occur in a relatively short time scale.
Human analysis nequires approximately 3 seconds per side of each fillet. An automated system with a throughput of one fillet per second is desirable for economic substitution.
The liunitation to providing high resolution imaging at human equivalent or even closer detail is the shear volume of data such an operation would create. A
high resolution image of a whole fillet would be too large for processing and analysis in a useful time scale. A fillet of an estimated 200 mm x 400 mm, at human equivalent resolution of 5 lines per mm, would generate an array of approximately 3000 x pixels. The sensors necessary to pick up such a large array are very expensive. A system capable of capturing and processing data at the necessary rate cannot be economically irnplemented. Emulation of human eye performance cannot be achieved using brute force.
To achieve human equivalent resolution, it has been proposed in the present irivention to mimic human visual processing. The human eye has a small central color sensitive area made up of cones capable of very high resolution. The majority of visual sensors which comprise the peripheral vision system are lower resolution monochrome rods. In human visual inspection, a broad scan by the peripheral vision system uses the attention mechanism to provide an initial identification of anomalies. The central vision sYstem is then used to examine anomalies in detail for final identification.
Eye niovement, known as saccadic fixation, jumps to relocate the eye position approximately 3-10 times per second having the effect of assembling a composite view of successive high resolution images. Imitating this two stage vision hierarchy involves collection of low resolution scene information over a wide area (typically 0.6 radian solid angle), providing an attention mechanism and directing a narrow field of view (about 17 mrads at a resolution of about 0.1 mrad) for close inspection of features of interest to assemble substantially complete high resolution information. Human eye saccadic fixation rate at about 10Hz has been emulated in a camera orienting system known as Agile Eye.
Flowever, the selective acquisition of high resolution information at relatively high speed (30-60 Hz) has not previously been achieved.
The present invention utilizing the post-objective scanning design to provide saccadic emulation, imitates human visual acquisition at an order of magnitude faster.
By achieving a 30-60 Hz acquisition rate, data acquisition can be matched to the normal fi-ame rate of a video camera, as commonly available, to capture 30 frames or 60 fields per second. Thus a relatively simple and inexpensive system is optimized to view an entirely different scene with each frame or field.
SUMMARY OF THE INVENTION
The present invention has found that significant savings in time can be achieved using the rapid and precise positioning capability of a two axis post-objective scanner in cooperation with a visual image sensor. Physical positioning time is reduced, and only selected high resolution data is acquired for processing. A video camera with a narrow field of view can be directed selectively to acquire only necessary target data, without acquiring high resolution image data of the complete object or scanning past the complete object thus reducing the total volume of data which must be handled. Preferably, the system according to the present invention is used in combination with an additional video camera having a wide field of view and an initial attention mechanism, such as blob analysis or other known image processing routines, to generate locations for closer inspection by the high resolution system.
Accordingly, the present invention provides an apparatus for the inspection of objects in an inspection field for the presence of a target, comprising:
a light source to illuminate the object;
first and second optic sensor means each being directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor means for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor; and second processor means associated with the second optic sensor means for generating an output of the target image data for analysis.
Desirably, the positioning means associated with the second optic sensor means comprises a post-objective scanner, the positioning means associated with the second optic sensor means comprises a post-objective scanner, the target includes an inspection field having a plane, the post-objective scanner is selectively directable in two axes toward a location on the plane in the inspection field, the post-objective scanner is directable sequentially towards each target location, the post-objective scanner has a random access of less than 33 milliseconds, and the post-objective scanner has a random access of about 3 milliseconds.
In a preferred embodiment, the present invention further includes a wide angle optic sensor means to acquire image data of the object to provide to the first processor means for detecting any target and for generating location data for each target.
In a preferred embodiment, the present invention further includes a high resolution optic sensor means having a smaller field of view than that of the wide angle sensor means, wherein the high resolution image includes a target substantially in the absence of unnecessary data.
Preferably, the wide angle sensor means and the high resolution sensor means are charge coupled device video cameras, the target comprises a visual anomaly, the first processor means includes a blob analysis routine for detecting visual anomalies, the wide angle optic sensor means and the high resolution optic sensor means comprise the same apparatus operating in different modes, and the second optic sensor means includes a field of view and the field of view of the high resolution optic sensor means is variable in response to the target size.
The present invention is preferably for the inspection of defects, including parasites in fish fillets.
In a further embodiment, the present invention provides an apparatus for the inspection of objects for the presence of targets including an optic sensor, the improvement comprising positioning means for saccadic emulation associated with the optic sensor for directing image data of a target to the optic sensor having a post-objective scanner.
It is further preferable, the target includes an inspection field having a plane and the post-objective scanner is selectively directable in two axes towards any location on the plane in the inspection field, the post-objective scanner has a random access of less than 33 milliseconds, the post-objective scanner has a random access of about 3 milliseconds, the optic sensor comprises a charge coupled device video camera having a frame rate and the post-objective scanner has a random access rate which can be synchronized with the frame rate of the camera, and further includes a processor means including a blob analysis routine.
In an additional embodiment, the present invention comprises an apparatus for the inspection of objects for visual targets, comprising:
a light source;
a first video imaging means for acquiring a visual image having a field of view including at least a portion of an object and for transmitting image data to a central processor;
a second video imaging means for acquiring a second visual image having a field of view smaller than the field of view of the first video imaging means and for transmitting image data to a central processor;
a central processor for receiving image data from the first and second video imaging means and for generating an output for analysis including:
a frame grabber for converting the image data of first and second video imaging means to digital data and isolating a single image;
a digital signal processor assembly programmed for morphological manipulation of the image data from the first video imaging means to detect individual visual anomalies as targets and for generating location coordinates for each target; and a post-objective scanner associated with the second video imaging means including:
a mirror directable in two axes for selectively directing image data from a specified target location to the second video imaging means; and a driver for accessing location coordinates and directing the mirror towards a target location.
Desirably, the video imaging means comprises charge coupled device video cameras, the cameras having a frame rate, and the post-objective scanner has a random access rate which can be synchronized with the frame rate of the camera.
In another embodiment of the present invention comprises an automated station for the detection of defects including parasites in fish fillets, including:
a viewing head comprising:
first image recording means for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target; and second processor means for generating an output from the target image data for analysis.
A preferred method, according to the present invention, for the inspection of objects for the presence of targets comprises:
detecting any targets on the object;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target; and generating an output for analysis.
Preferably, the method according to the present invention further includes synchronizing the saccadic movement emulation to sequentially capture a high resolution image of each target detected at the frame rate of a video imaging means, and the method includes a first step of treating an object with a marker prior to inspection to identify any targets and a detection means is provided for cooperation with the selected marker for detecting the presence of a target.
A further preferred method according to the present invention comprises a method for the inspection of fish fillets for the presence of defects including parasites, comprising the steps of:
9a capturing an image of at least a portion of a fillet;
detecting any defects or potential parasites as targets;
generating a location for a detected target;
emulating saccadic movement to selectively capture a higher resolution image of a target;
processing target image data to determine whether a target is a defect or parasite; and identifying a fillet having a defect or parasite.
In another embodiment of the present invention there is provided an apparatus for the inspection of objects in an inspection field for the presence of a target comprising:
a light source;
first and second optic sensor means each directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor means whereby the high resolution target image data is acquired and processed generally immediately following the generation of target location data by the first optic sensor and the first processor means; and second processor means associated with the second optic sensor for generating an output of the target image data for analysis.
Preferably, the positioning means associated with the optic sensor means comprises a post-objective scanner, the post-objective scanner is selectively directable in two axes toward a location on a plane in the inspection field, is directable sequentially towards each target location, and has a random access of less than 33 milliseconds.
It is further preferable the scanner has a random access of about 3 milliseconds.
9b In the above embodiment it is preferable the first optic sensor means is a wide angle optic sensor means provided to acquire image data of the object to provide to the first processor means for detecting any target and for generating location data for each target, the wide angle optic sensor means captures low resolution image data of the object, and the second optic sensor means comprises a high resolution optic sensor means having a smaller field of view than that of the wide angle sensor means, the wide angle sensor means and the high resolution sensor means are charge coupled device video cameras, and the targets comprise visual anomalies.
Moreover, the first processor means desirably includes a blob analysis routine for detecting the visual anomalies.
In addition, it is preferred in the above embodiment, the wide angle optic sensor means and the high resolution optic sensor means comprise the same apparatus operating in different modes, the high resolution optic sensor means includes a field of view and the field of view of the high resolution optic sensor means is variable in response to the target size, and the high resolution image includes a target substantially in the absence of unnecessary data.
In another embodiment of the present invention there is provided a method for the inspection of objects for the presence of targets comprising the steps of providing an apparatus as defined above;
detecting any targets on the object with the first optic sensor means;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target with the second sensor means generally immediately following the step of generating location data and without moving the object; and, generating an output for analysis.
Preferably, saccadic movement emulation step comprises a video imaging means, having a frame rate, and when there is a sequence of targets detected the capture of the high resolution image of each target detected is sequentially synchronized to the frame rate of a 9c video imaging means, and the method includes a first step of treating an object with a marker prior to inspection to identify any targets and a detection means is provided for cooperation with the selected marker for detecting the presence of a target, or the method includes a first step of capturing a first image of at least a portion of an object from which any targets can be detected.
Further, it is desirable the above embodiment includes identifying visual anomalies as targets for detection.
Moreover, it is desirable the object comprises a fish fillet, and the targets include parasites, the step of generating an output for analysis includes the step of processing target image data to determine whether the target is a defect or a parasite and identifying a fillet as having a defect or parasite.
In another embodiment of the present invention, there is provided an automated station for the detection of defects including parasites in fish fillets including:
a viewing head comprising:
first image recording means for viewing the object concurrently with the first image recording means and for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target;
9d second processor means for generating an output from the target image data for analysis; and wherein the first and second image recording means, and the first and second processor means operate generally one immediately following the other to emulate the functioning of a human eye to generally nearly simultaneously or in rapid succession identify targets within a broad field of view and resolve such targets within a narrow field of view at higher resolution.
It is an advantage of the present invention to provide the capability to detect targets, and to be able to selectively isolate targets for detailed examination. Speed is gained according to the present invention by acquiring only necessary high resolution data for processing and analysis. The system advantageously reduces the amount of data to be processed and consequently the time required. Post-objective scanning permits the camera to remain stationary as only the mirrors need to be moved. As a result, the speed is increased to accurately enable image acquisition synchronous with the frame rate of a standard video camera, and the driving power is reduced.
It is a further advantage to achieve fast and accurate performance utilizing commonly available, and hence cost effective, hardware.
The present invention is suitable for many visual examination tasks, particularly where multiple target acquisition is necessary, or in the examination of a relatively large field in a short time period. This system can be adapted for use in other examination tasks where identification in complex and irregular scenes must be made, such as other areas of food processing, inspection of natural materials, and medical diagnostics.
BRIEF DESCRIPTION OF THE FIGURES
The invention will be readily understood by the following description of a preferred embodiment with reference to the following drawings in which:
Figure 1 is a schematic illustration of a bench scale embodiment of the present invention;
Figure 2 is a cut-away view of the scan head of Figure 1;
Figure 3 is a schematic illustration of the embodiment of Figure 1 showing the logical process path;
Figure 4 is a schematic illustration of an embodiment of the present invention for the inspection of fish fillets; and Figure 5 is a flow chart schematically illustrating the first and second processor routines.
Like numerals are used throughout to designate like elements.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
An embodiment of the invention as a system is shown generally as 10 in Figure 1. As seen in a schematic bench scale model, a scan head 100 is adjustably mounted on a standard 101. The object 12 having test targets 14 is presented to the cameras 20, 30 (see Fig. 2) in the scan head 100, on table 16. The cameras 20, 30 are preferably video imaging means, such as CCD (charge coupled device) video cameras. Other equivalent image recording means may be substituted. A candling table 16 is provided for back-lighting a fillet 12 in the case of fish inspection. The image data acquired by the cameras can be displayed on the monitor 28, or in real time on the video monitor 29. General lighting 18 is also provided.
Lighting requirements, as known in the art, will change for each application. In this configuration the object 12 is presented manually and remains in the same position for image acquisition by the low and high resolution cameras 20, 30. The system 10 can also operate in a dynamic mode (seen in Figure 3) with the objects 12 travelling on an continuous motion conveyor 116 at a fixed rate.
As represented in Figure 3, the first camera 20 has a wide field of view Flwhich includes the entire object 12. A second camera 30 with a high magnification lens 31 has a narrow field of view FZ including only a reduced portion of the object 12 large enough to capture a potential target 14 at higher resolution substantially in the absence of unnecessary data. The narrow field of view F2 can be adjusted to accommodate the anticipated target size. If the speed of operation is reduced, the field of view FZ can be variable, for ivnstance using an automatic zoom lens, in response to the actual target size 14. Similarly, under reduced speed operation, a single camera can be adapted to work in low resolut;ion and high resolution modes using an automatic zoom lens to change the field of view.
A positioning device 32 comprises a high speed two-axis post-objective optical scanner for directing the field of view of the second camera 30. Seen in detail in Figure 2, the scan head 100 houses a first low resolution camera 20 and lens 21, and a high resolution carnera 30 which cooperates with a high magnification lens 31, an x-y scanner 32 having a niirror 34 on a mount 35 for two axis rotation.
Associated with the first CCD camera 20 for processing controlled by a CPU 26, seen schematically in Figure 3, is a frame grabber 22 for isolating a single digital image and a target locator 24 for detecting visual targets and for generating x-y coordinate target location information. Coordinate location information is provided to a scanner driver 36 for directing the field of view FZ of the second camera 30 towards a target. At 30-60Hz the scanner 32 directs new image data isolated by the frame grabber 22 at the firame rate of the camera 30. The CPU 26 also provides synchronization for the lighting 16, 18, cameiras 20, 30, and scanner 32. Processing operations associated with each camera are shown as separate loops, however the necessary hardware is actually common between theni.
Processing in the CPU 26 generates target location data, depending on the application, from a data table in combination with position registration information from a first sensor, a subtraction operation, location or image information from a first sensor.
The first sensor is provided to identify the presence of an object and the location of targets. Since the initial identification is not for visual analysis, other sensors, as available in the art, may be used to generate location information. Targets maybe pre-marked for detection with appropriate detection means, for instance by florescence in combination with a photodetector, or other well known markers. Advantageously target locations are icientified froin a first low resolution image acquired by a visual sensor, such as a line scan or CCD camera, for the inspection of irregular objects with a blob analysis routine, or comparable pattern recognition operation to detect targets and generate locations for detailed inspection. A frame grabber system 22 is required which can perform image digitization and image processing within one second. A programmable DSP
(digital signal processor) board or board with multiple DSPs 24, for blob analysis or other known irnage processing routines, may be interfaced with the frame grabber 22, for instance, to achieve the p:rocessing speed required to detect targets.
A second CCD camera 30 and high magnification lens 31 acquire isolated target irnage data, without unnecessary high resolution data. The scan system 32, 34, 36, allows the high resolution camera 30 to selectively obtain visual data from any location on the surface of the: fillet 12 without taking a high resolution image of the whole object 12.
The scanner 32 has a mirror 34 with two axis control at very rapid access time to direct light back to the high resolution camera 30. The scanner 32 and mirror 34 effectively position the fi"eld of view of the high resolution camera 30 anywhere in the object field.
In operation the first camera 20 acquires a low resolution image which is processed only once per object 12 to detect any targets 14. A frame grabber 22 isolates a single frame and converts the visual data to digital data. Processing in the central processor 26 using a blob analysis routine or equivalent target locator 24 then generates coordinates indicating the location of the targets 14. Coordinate data is used to drive the scanner 32, and mirror 34 sequentially to each location where the second camera 30 obtains high resolution image data of each target 14 within a narrow field of view. The h:igh resolution target image data can then be processed for further analysis, such as a ninning count, comparison, categorizing, acceptance or rejection, etc. The presence of known features may be subtracted leaving only foreign anomalies for inspection.
Visually targets are identified by shape, color, contrast, texture or other changes iri appearance which may be highlighted by lighting techniques such as lighting angle or back lighting, selected colored light, edge detection, misalignment, surface irregularity, color value, as, known in the art. Processing visual data to detect targets may be specified to certain pai-ameters or compared to a data library using algorithmic morphological rnanipulation such as blob analysis, thresholding, edge detection, dilation, erosion, subtraction, inter alia known in the art, which will vary depending on the specific application. 'The benefit in more specific target detection is that fewer target images niust be acquiired and processed, thus increasing throughput speed.
Figure 5 illustrates the first processor routine 300, and the second processor routine 400 in a fillet inspection application. In this preferred operation, as seen in 400, each target is inspected sequentially. A final output is made to accept or reject the fillet.
The system then returns to wait for the next fillet.
As seen in Figure 4 an embodiment including the present invention for fish irispection 200 includes a feed conveyor 216 adapted to stop at the inspection station or to advance the fillet 12 through at a controlled rate. Fillets 12 may be supplied from an additional conveyor upstream (not shown), or by manual placement. The inspection station comprises a candling light source 218, a scan head 100 (housed in a protective upper cabinet 219 shrouding the inspection field, and preferably separated by glass from the fish handling environment) and a computerized control interface 228. Once the inspection is complete the fillet 12 is advanced by the conveyor 216 and classified according to the results. This is represented by a mechanical deflector 220 and separate containers 222, 223. Separate outfeed conveyors (not shown) may alternatively be provided to convey the separated fillets for further processing according to the inspection results.
"ne scan head 100 for parasite detection uses two low resolution (eg. 512 x p:ixel) CCD viideo cameras, preferably having 24 bit color capability. For the present application, the lowest acceptable resolution to detect all potential targets in the low resolution camera 20 is 550 m or 0.5 lines per mm, for instance using a 6 mm lens. In this application a resolution of 78 m or 5 lines per mm is needed for reliable d:iscrimination. of features using a 60-300 mm lens 31 with the high resolution camera 30.
This provides a first field of view of approximately 200 x 400 mm and a narrow field of view of approximately 10-40 mm square. A scanner controllable over 20 at 3 msec.
access time is used in the present embodiment. The mirror 34 is 10mm, selected to cooperate with the aperture of the high resolution camera 30. Intense illumination is provided to compensate for a relatively small mirror which reduces drive power requirements and increases positioning speed. In the fillet inspection application approximately 20-30 targets per fillet must be inspected at high resolution in the one second throughput time. The elements used in this application have been selected as commercially available components for cost control. As apparent to a person sldlled in the art, equivalent components can be substituted for specific applications as needed.
The device according to the present invention has also been adapted for the icientification of visual blemishes in other applications, such as paint and other surface treatments in manufactured articles. Changes in color, texture, contrast and shape are easily identified. For large articles a series of inspections may be needed to inspect the complete object.
The present invention creates an output which can be used as a culling mechanism to reduce the number of objects for human inspection, as a sorting system by numerical tlireshold, as a search system for features in a visual data set, for off-site human irispection or with a pattern recognition and analysis system for further automated processing. Other applications of the invention as defined in the appended claims will be apparent to persons of skill in the art.
Claims (51)
1. An apparatus for the inspection of objects in an inspection field for the presence of a target comprising:
a light source to illuminate the object;
first and second optic sensor means each directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor means for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor means; and second processor means associated with the second optic sensor means for generating an output of the target image data for analysis.
a light source to illuminate the object;
first and second optic sensor means each directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor means for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor means; and second processor means associated with the second optic sensor means for generating an output of the target image data for analysis.
2. The apparatus for the inspection of objects as defined in claim 1, wherein the positioning means associated with the second optic sensor means comprises a post-objective scanner.
3. The apparatus for the inspection of objects as defined in claim 2, wherein the target includes an inspection field having a plane and the post-objective scanner is selectively directable in two axes toward a location on the plane in the inspection field.
4. The apparatus for the inspection of objects as defined in claim 2 or 3, wherein the post-objective scanner is directable sequentially towards each target location.
5. The apparatus for the inspection of objects as defined in any one of claims 2 to 4, wherein the post-objective scanner has a random access of less than 33 milliseconds.
6. The apparatus for the inspection of objects as defined in any one of claims 2 to 5, wherein the post-objective scanner has a random access of about 3 milliseconds.
7. The apparatus for the inspection of objects as defined in claim 1, wherein the first optic sensor means includes a wide angle optic sensor means provided to acquire image data of the object to provide to the first processor means for detecting any target and for generating location data for each target.
8. The apparatus for the inspection of objects as defined in claim 7, wherein the first optic sensor means captures low resolution image data of the object, and the second optic sensor means comprises a high resolution optic sensor means having a smaller field of view than that of the wide angle optic sensor means.
9. The apparatus for the inspection of objects as defined in any one of claims 1 to 8, wherein the high resolution image includes a target substantially in the absence of unnecessary data.
10. The apparatus for the inspection of objects as defined in claim 7, wherein the wide angle sensor means and the high resolution sensor means are charge coupled device video cameras.
11. The apparatus for the inspection of objects as defined in any one of claims 1 to 10, wherein the target comprises a visual anomaly.
12. The apparatus for the inspection of objects as defined in any one of claims 1 to 11, wherein the first processor means includes a blob analysis routine for detecting visual anomalies.
13. The apparatus for the inspection of objects as defined in claim 7, wherein the wide angle optic sensor means and the high resolution optic sensor means comprise the same apparatus operating in different modes.
14. The apparatus for the inspection of objects as defined in claim 8, wherein the second optic sensor means includes a field of view and the field of view of the high resolution optic sensor means is variable in response to the target size.
15. Use of the apparatus for the inspection of objects as defined in claim 1 for the detection of defects including parasites in fish fillets.
16. In an apparatus for the inspection of objects for the presence of targets including an optic sensor, the improvement comprising:
positioning means for saccadic emulation associated with an optic sensor for directing image data of a target to the optic sensor having a post-objective scanner.
positioning means for saccadic emulation associated with an optic sensor for directing image data of a target to the optic sensor having a post-objective scanner.
17. The apparatus for the inspection of objects as defined in claim 16, wherein the target includes an inspection field having a plane and the post-objective scanner is selectively directable in two axes towards any location on the plane in the inspection field.
18. The apparatus for the inspection of objects as defined in claim 16 or 17, wherein the post-objective scanner has a random access of less than 33 milliseconds.
19. The apparatus for the inspection of objects as defined in any one of claims 16 to 18, wherein the post-objective scanner has a random access of about 3 milliseconds.
20. The apparatus for the inspection of objects as defined in any one of claims 16 to 19, wherein the optic sensor comprises a charge coupled device video camera having a frame rate and the post-objective scanner has a random access rate which can be synchronized with the frame rate of the camera.
21. The apparatus for the inspection of objects as defined in any one of claims 16 to 20, further including a processor means including a blob analysis routine.
22. An apparatus for the inspection of objects for visual targets comprising:
a light source;
a first video imaging means for acquiring a visual image having a field of view including at least a portion of an object and for transmitting image data to a central processor;
a second video imaging means for acquiring a second visual image having a field of view smaller than the field of view of the first video imaging means and for transmitting image data to a central processor;
a central processor for receiving image data from the first and second video imaging means and for generating an output for analysis including:
a frame grabber for converting the image data of first and second video imaging means to digital data and isolating a single image;
a digital signal processor assembly programmed for morphological manipulation of the image data from the first video imaging means to detect individual visual anomalies as targets and for generating location coordinates for each target; and a post-objective scanner associated with the second video imaging means including:
a mirror directable in two axes for selectively directing image data from a specified target location to the second video imaging means, and a driver for accessing location coordinates and directing the mirror towards a target location.
a light source;
a first video imaging means for acquiring a visual image having a field of view including at least a portion of an object and for transmitting image data to a central processor;
a second video imaging means for acquiring a second visual image having a field of view smaller than the field of view of the first video imaging means and for transmitting image data to a central processor;
a central processor for receiving image data from the first and second video imaging means and for generating an output for analysis including:
a frame grabber for converting the image data of first and second video imaging means to digital data and isolating a single image;
a digital signal processor assembly programmed for morphological manipulation of the image data from the first video imaging means to detect individual visual anomalies as targets and for generating location coordinates for each target; and a post-objective scanner associated with the second video imaging means including:
a mirror directable in two axes for selectively directing image data from a specified target location to the second video imaging means, and a driver for accessing location coordinates and directing the mirror towards a target location.
23. The apparatus for the inspection of objects as defined in claim 22, wherein the video imaging means comprises charge coupled device video cameras, the cameras having a frame rate, and the post-objective scanner has a random access rate which can be synchronized with the frame rate of the camera.
24. An automated station for the detection of defects including parasites in fish fillets including:
a viewing head comprising:
first image recording means for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target; and second processor means for generating an output from the target image data for analysis.
a viewing head comprising:
first image recording means for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target; and second processor means for generating an output from the target image data for analysis.
25. A method for the inspection of objects for the presence of targets comprising the steps of:
detecting any targets on the object;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target; and generating an output for analysis.
detecting any targets on the object;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target; and generating an output for analysis.
26. A method for the inspection, of objects as defined in claim 25, wherein the saccadic movement emulation step comprises a video imaging means, having a frame rate, and when there is a sequence of targets detected the capture of the high resolution image of each target detected is sequentially synchronized to the frame rate of the video imaging means.
27. A method as defined in claim 25 or 26, wherein the method includes a first step of treating an object with a marker prior to inspection to identify any targets and a detection means is provided for cooperation with the selected marker for detecting the presence of a target.
28. A method for the inspection of objects as defined in any one of claims 25 to 27, wherein the method includes a first step of capturing a first image of at least a portion of an object from which any targets can be detected.
29. A method for the inspection of objects as defined in claim 28, further including identifying visual anomalies as targets for detection.
30. A method for the inspection of fish fillets for the presence of defects including parasites comprising the steps of:
capturing an image of at least a portion of a fillet;
detecting any defects or potential parasites as targets;
generating a location for a detected target;
emulating saccadic movement to selectively capture a higher resolution image of a target;
processing target image data to determine whether a target is a defect or parasite; and identifying a fillet having a defect or parasite.
capturing an image of at least a portion of a fillet;
detecting any defects or potential parasites as targets;
generating a location for a detected target;
emulating saccadic movement to selectively capture a higher resolution image of a target;
processing target image data to determine whether a target is a defect or parasite; and identifying a fillet having a defect or parasite.
31. An apparatus for the inspection of objects in an inspection field for the presence of a target comprising:
a light source;
first and second optic sensor means each directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor means whereby the high resolution target image data is acquired and processed generally immediately following the generation of target location data by the first optic sensor and the first processor means; and second processor means associated with the second optic sensor for generating an output of the target image data for analysis.
a light source;
first and second optic sensor means each directed towards the object for concurrent viewing of the object;
first processor means associated with the first optic sensor for detecting a target and for generating location data for a target;
the second optic sensor means for selectively acquiring high resolution image data of a located target;
positioning means for saccadic emulation associated with the second optic sensor means for directing the image data of a target location to the second optic sensor means whereby the high resolution target image data is acquired and processed generally immediately following the generation of target location data by the first optic sensor and the first processor means; and second processor means associated with the second optic sensor for generating an output of the target image data for analysis.
32. An apparatus for the inspection of objects as defined in claim 31, wherein the positioning means associated with the optic sensor means comprises a post-objective scanner.
33. An apparatus for the inspection of objects as defined in claim 32, wherein the post-objective scanner is selectively directable in two axes toward a location on a plane in the inspection field.
34. An apparatus for the inspection of objects as defined in claim 32 or 33, wherein the post-objective scanner is directable sequentially towards each target location.
35. The apparatus for the inspection of objects as defined in any one of claims 32 to 34, wherein the post-objective scanner has a random access of less than 33 milliseconds.
36. The apparatus for the inspection of objects as defined in any one of claims 32 to 35, wherein the scanner has a random access of about 3 milliseconds.
37. An apparatus for the inspection of objects as defined in claim 31, wherein the first optic sensor means is a wide angle optic sensor means provided to acquire image data of the object to provide to the first processor means for detecting any target and for generating location data for each target.
38. An apparatus for the inspection of objects as defined in claim 37, wherein the wide angle optic sensor means captures low resolution image data of the object, and the second optic sensor means comprises a high resolution optic sensor means having a smaller field of view than that of the wide angle sensor means.
39. An apparatus for the inspection of objects as defined in claim 38, wherein the wide angle sensor means and the high resolution sensor means are charge coupled device video cameras.
40. An apparatus for the inspection of objects as defined in claim 37, wherein the targets comprise visual anomalies.
41. An apparatus for the inspection of objects as defined in claim 40, wherein the first processor means includes a blob analysis routine for detecting the visual anomalies.
42. The apparatus for the inspection of objects as defined in any one of claims 38 to 41, wherein the wide angle optic sensor means and the high resolution optic sensor means comprise the same apparatus operating in different modes.
43. The apparatus for the inspection of objects as defined in any one of claims 38 to 42, wherein the high resolution optic sensor means includes a field of view and the field of view of the high resolution optic sensor means is variable in response to the target size.
44. An apparatus for the inspection of objects as defined in any one of claims 38 to 43, wherein the high resolution image includes a target substantially in the absence of unnecessary data.
45. A method for the inspection of objects for the presence of targets comprising the steps of:
providing an apparatus as defined in claim 31;
detecting any targets on the object with the first optic sensor means;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target with the second sensor means generally immediately following the step of generating location data and without moving the object; and, generating an output for analysis.
providing an apparatus as defined in claim 31;
detecting any targets on the object with the first optic sensor means;
generating location data for a detected target;
emulating saccadic movement to selectively capture a high resolution image of a target with the second sensor means generally immediately following the step of generating location data and without moving the object; and, generating an output for analysis.
46. A method for the inspection of objects as defined in claim 45, wherein the saccadic movement emulation step comprises a video imaging means, having a frame rate, and when a sequence of targets is detected the capture of the high resolution image of each target detected is sequentially synchronized to the frame rate of the video imaging means.
47. A method as defined in claim 45 or 46, wherein the method includes a first step of treating an object with a marker prior to inspection to identify any targets and a detection means is provided for cooperation with the selected marker for detecting the presence of a target.
48. A method for the inspection of objects as defined in any one of claims 45 to 47, wherein the method includes a first step of capturing a first image of at least a portion of an object from which any targets can be detected.
49. A method for the inspection of objects as defined in claim 48, further including identifying visual anomalies as targets for detection.
50. A method as defined in claim 45, wherein the object comprises a fish fillet, and the targets include parasites, the step of generating an output for analysis includes the step of processing target image data to determine whether the target is a defect or a parasite and identifying a fillet as having a defect or parasite.
51. An automated station for the detection of defects including parasites in fish fillets including:
a viewing head comprising:
first image recording means for viewing the object concurrently with the first image recording means and for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target;
second processor means for generating an output from the target image data for analysis; and wherein the first and second image recording means, and the first and second processor means operate generally one immediately following the other to emulate the functioning of a human eye to generally nearly simultaneously or in rapid succession identify targets within a broad field of view and resolve such targets within a narrow field of view at higher resolution.
a viewing head comprising:
first image recording means for viewing the object concurrently with the first image recording means and for acquiring image data of an object in an inspection field;
second image recording means having a smaller field of view than that of the first image recording means for selectively acquiring higher resolution image data;
positioning means associated with the second image recording means, for directing image data of substantially only a target to the second image recording means including a post-objective scanner directable in two axes toward any location on a plane in the inspection field;
a candling table for back-lighting a fillet;
a light source for providing illumination of the inspection field;
advancing means for presenting a fillet on the candling table to the first and second cameras;
a support for supporting the viewing head a distance from the candling table;
first processor means for processing the image data of the object to isolate any targets, and for generating location data for each target;
second processor means for generating an output from the target image data for analysis; and wherein the first and second image recording means, and the first and second processor means operate generally one immediately following the other to emulate the functioning of a human eye to generally nearly simultaneously or in rapid succession identify targets within a broad field of view and resolve such targets within a narrow field of view at higher resolution.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US08/927,332 US6061086A (en) | 1997-09-11 | 1997-09-11 | Apparatus and method for automated visual inspection of objects |
| CA002227815A CA2227815C (en) | 1997-09-11 | 1998-01-23 | Apparatus and method for automated visual inspection of objects |
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US08/927,332 US6061086A (en) | 1997-09-11 | 1997-09-11 | Apparatus and method for automated visual inspection of objects |
| US08/927,332 | 1997-09-11 | ||
| CA002227815A CA2227815C (en) | 1997-09-11 | 1998-01-23 | Apparatus and method for automated visual inspection of objects |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CA2227815A1 CA2227815A1 (en) | 1999-03-11 |
| CA2227815C true CA2227815C (en) | 2007-05-22 |
Family
ID=25679996
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CA002227815A Expired - Lifetime CA2227815C (en) | 1997-09-11 | 1998-01-23 | Apparatus and method for automated visual inspection of objects |
Country Status (2)
| Country | Link |
|---|---|
| US (1) | US6061086A (en) |
| CA (1) | CA2227815C (en) |
Families Citing this family (52)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2998791B2 (en) * | 1996-10-31 | 2000-01-11 | 日本電気株式会社 | 3D structure estimation device |
| US6856698B1 (en) * | 1997-11-26 | 2005-02-15 | Cognex Corporation | Fast high-accuracy multi-dimensional pattern localization |
| US7016539B1 (en) | 1998-07-13 | 2006-03-21 | Cognex Corporation | Method for fast, robust, multi-dimensional pattern recognition |
| US6630998B1 (en) | 1998-08-13 | 2003-10-07 | Acushnet Company | Apparatus and method for automated game ball inspection |
| US6864970B1 (en) * | 2000-10-11 | 2005-03-08 | Best N.V. | Apparatus and method for scanning products with a light beam to detect and remove impurities or irregularities in a conveyed stream of the products |
| WO2002031945A2 (en) * | 2000-10-13 | 2002-04-18 | Clarity, Llc | Magnetic actuation and positioning |
| US6985616B2 (en) * | 2001-10-18 | 2006-01-10 | Robodesign International, Inc. | Automated verification and inspection device for sequentially inspecting microscopic crystals |
| US6462812B1 (en) | 2000-11-21 | 2002-10-08 | Callaway Golf Company | Golf ball indicia verification system |
| IL149587A (en) * | 2001-05-11 | 2005-11-20 | Orbotech Ltd | Optical inspection system employing a staring array scanner |
| US6786096B2 (en) | 2001-11-28 | 2004-09-07 | Battelle Memorial Institute | System and technique for detecting the presence of foreign material |
| US6992771B2 (en) * | 2001-11-28 | 2006-01-31 | Battelle Memorial Institute | Systems and techniques for detecting the presence of foreign material |
| US6879082B2 (en) * | 2002-03-25 | 2005-04-12 | Clarity Technologies, Inc. | Electromagnetic positioning |
| US20030193562A1 (en) * | 2002-04-15 | 2003-10-16 | Corzillus Brian S. | Natural vision-based video surveillance system |
| US6879389B2 (en) * | 2002-06-03 | 2005-04-12 | Innoventor Engineering, Inc. | Methods and systems for small parts inspection |
| US7103208B2 (en) * | 2002-08-26 | 2006-09-05 | Eastman Kodak Company | Detecting and classifying blemishes on the transmissive surface of an image sensor package |
| US6880759B2 (en) * | 2003-05-23 | 2005-04-19 | Symagery Microsystems Inc. | Optical reader station |
| US7190834B2 (en) | 2003-07-22 | 2007-03-13 | Cognex Technology And Investment Corporation | Methods for finding and characterizing a deformed pattern in an image |
| US8081820B2 (en) | 2003-07-22 | 2011-12-20 | Cognex Technology And Investment Corporation | Method for partitioning a pattern into optimized sub-patterns |
| WO2005026919A2 (en) * | 2003-09-11 | 2005-03-24 | Bradford Addison Clough | Acquisition and analysis of time location-specific image data |
| US8437502B1 (en) | 2004-09-25 | 2013-05-07 | Cognex Technology And Investment Corporation | General pose refinement and tracking tool |
| EP1802968A1 (en) * | 2004-09-30 | 2007-07-04 | Paratest ApS | A method and a system for detection of trichinella larvae in meat samples |
| JP4654022B2 (en) * | 2004-12-24 | 2011-03-16 | 株式会社サキコーポレーション | Substrate visual inspection device |
| DE102005018855B4 (en) * | 2005-04-22 | 2010-01-28 | Theta System Elektronik Gmbh | Apparatus for inspecting printed products |
| DE102005029901B4 (en) * | 2005-06-25 | 2022-10-06 | Modi Modular Digits Gmbh | Device and method for visually detecting two-dimensional or three-dimensional objects |
| US20070106161A1 (en) * | 2005-10-19 | 2007-05-10 | The Regents Of The University Of California | Medical diagnostic imaging using visible light |
| WO2008016309A1 (en) * | 2006-08-04 | 2008-02-07 | Sinvent As | Multi-modal machine-vision quality inspection of food products |
| US8661671B2 (en) | 2006-09-12 | 2014-03-04 | Benteler Automotive Corporation | Method for making catalytic converters with automated substrate crack detection |
| US7992851B2 (en) * | 2007-08-01 | 2011-08-09 | Acushnet Company | Device for automatic indexing of a golf ball |
| US7387070B1 (en) | 2007-08-01 | 2008-06-17 | Acushnet Company | Method and apparatus for automatic indexing of a golf ball |
| GB0717019D0 (en) * | 2007-09-01 | 2007-10-10 | Rixen Uk Ltd | Audiovisual Terminal |
| JP2009168746A (en) * | 2008-01-18 | 2009-07-30 | Sumitomo Electric Ind Ltd | Inspection method and inspection device |
| EP2325625A1 (en) * | 2008-08-07 | 2011-05-25 | Kde Corporation | Inspection system |
| US20100091094A1 (en) * | 2008-10-14 | 2010-04-15 | Marek Sekowski | Mechanism for Directing a Three-Dimensional Camera System |
| US20110063446A1 (en) * | 2009-09-14 | 2011-03-17 | Mcmordie David | Saccadic dual-resolution video analytics camera |
| CL2009002085A1 (en) | 2009-11-16 | 2011-03-11 | Univ Pontificia Catolica Chile | Method and system to automatically analyze in real time the quality of samples of fish meats that circulate on a conveyor belt, which allow surface defects to be detected and classify meats according to quality standards, based on the segmentation of the captured images. |
| NO331345B1 (en) | 2010-02-05 | 2011-12-05 | Esben Beck | Device and method for damaging parasites on fish |
| WO2012168322A2 (en) | 2011-06-06 | 2012-12-13 | 3Shape A/S | Dual-resolution 3d scanner |
| EE201200005A (en) | 2012-03-27 | 2013-12-16 | Humal Priit | Apparatus for diagnosing or controlling bee varroasis, method and software for parasite detection |
| CN103297661A (en) * | 2013-06-04 | 2013-09-11 | 四川艾普视达数码科技有限公司 | Infrared monitoring device with fixed-point thinning camera shooting function |
| US9679224B2 (en) | 2013-06-28 | 2017-06-13 | Cognex Corporation | Semi-supervised method for training multiple pattern recognition and registration tool models |
| KR102145960B1 (en) * | 2013-07-19 | 2020-08-20 | 삼성디스플레이 주식회사 | Apparatus and method for inspecting sealing |
| WO2015079268A1 (en) * | 2013-11-28 | 2015-06-04 | Femtonics Kft. | Optical microscope system for simultaneous observation of spatially distinct regions of interest |
| ES2552405B1 (en) * | 2014-05-27 | 2016-09-14 | Tecnología Marina Ximo, S.L. | System and method of detection of Anisakis parasites in fish fillets |
| US20170004344A1 (en) * | 2015-07-02 | 2017-01-05 | Canon Kabushiki Kaisha | Robust Eye Tracking for Scanning Laser Ophthalmoscope |
| US9870503B2 (en) * | 2015-12-31 | 2018-01-16 | International Business Machines Corporation | Visual object and event detection and prediction system using saccades |
| CN105787460B (en) * | 2016-03-14 | 2019-09-10 | 南京市疾病预防控制中心 | A kind of schistosoma miracidium automatic recognition system |
| US10345789B2 (en) | 2016-06-21 | 2019-07-09 | Scientific Games International, Inc. | System and method for variable perforation profiles in a stack of lottery tickets |
| CN107727232B (en) * | 2017-10-24 | 2023-07-04 | 中国科学院上海技术物理研究所 | Geometric registration testing device and method |
| US11825814B2 (en) | 2017-12-20 | 2023-11-28 | Intervet Inc. | System for external fish parasite monitoring in aquaculture |
| US11632939B2 (en) * | 2017-12-20 | 2023-04-25 | Intervet Inc. | System for external fish parasite monitoring in aquaculture |
| US10869489B2 (en) | 2018-08-31 | 2020-12-22 | John Bean Technologies Corporation | Portioning accuracy analysis |
| CZ2020406A3 (en) * | 2020-07-10 | 2021-09-08 | České vysoké učenà technické v Praze | Mobile electronic device for indicating bee diseases and hive conditions |
Family Cites Families (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US3704342A (en) * | 1970-01-22 | 1972-11-28 | Dynarad | Infrared scanning system |
| JPS61229171A (en) * | 1985-04-04 | 1986-10-13 | Bridgestone Corp | Method and device for reading tire information |
| US4744131A (en) * | 1985-07-12 | 1988-05-17 | Nordischer Maschinenbau Rud. Baader Gmbh+C Kg | Apparatus for handling fish fillets for the purpose of quality inspection |
| JPS62173731A (en) * | 1986-01-28 | 1987-07-30 | Toshiba Corp | Inspection device for surface of article to be inspected |
| DK157380C (en) * | 1986-11-06 | 1991-08-12 | Lumetech As | METHODS OF OPTICAL, BODY-FREE MEASUREMENT OF MEAT TEXTURE |
| US4872052A (en) * | 1986-12-03 | 1989-10-03 | View Engineering, Inc. | Semiconductor device inspection system |
| US5051825A (en) * | 1989-04-07 | 1991-09-24 | Pressco, Inc. | Dual image video inspection apparatus |
| CA1251863A (en) * | 1988-02-29 | 1989-03-28 | Kevin Mccarthy | Fish sorting machine |
| US5173796A (en) * | 1991-05-20 | 1992-12-22 | Palm Steven G | Three dimensional scanning system |
| US5255114A (en) * | 1991-11-12 | 1993-10-19 | Eastman Kodak Company | High resolution scanner |
| DE69329820T2 (en) * | 1992-04-13 | 2001-06-07 | Australian Meat & Live Stock | MEAT IMAGE ANALYSIS. |
| JPH05313085A (en) * | 1992-05-08 | 1993-11-26 | Komatsu Ltd | Scanner |
| US5257101A (en) * | 1992-06-01 | 1993-10-26 | Board Of Regents Of The University Of Ok | Apparatus and method for selecting a log orientation and pattern for sawing |
| JP3212389B2 (en) * | 1992-10-26 | 2001-09-25 | 株式会社キリンテクノシステム | Inspection method for foreign substances on solids |
| US5341181A (en) * | 1992-11-20 | 1994-08-23 | Godard Roger R | Systems and methods for capturing and presentng visual information |
| JPH09500963A (en) * | 1993-07-22 | 1997-01-28 | ミネソタ マイニング アンド マニュファクチャリング カンパニー | Three-dimensional spatial calibration method and apparatus for machine vision applications |
| US5517235A (en) * | 1993-11-03 | 1996-05-14 | Control Automation, Inc. | Method and apparatus for inspecting printed circuit boards at different magnifications |
| US5784484A (en) * | 1995-03-30 | 1998-07-21 | Nec Corporation | Device for inspecting printed wiring boards at different resolutions |
-
1997
- 1997-09-11 US US08/927,332 patent/US6061086A/en not_active Expired - Lifetime
-
1998
- 1998-01-23 CA CA002227815A patent/CA2227815C/en not_active Expired - Lifetime
Also Published As
| Publication number | Publication date |
|---|---|
| US6061086A (en) | 2000-05-09 |
| CA2227815A1 (en) | 1999-03-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2227815C (en) | Apparatus and method for automated visual inspection of objects | |
| EP1151283B1 (en) | Cylindrical object surface inspection system | |
| CN209764751U (en) | Surface defect detection system | |
| US5926268A (en) | System and method for stress detection in a molded container | |
| EP0493487B1 (en) | Dual image video inspection apparatus | |
| US5095204A (en) | Machine vision inspection system and method for transparent containers | |
| CN105259189B (en) | The defect imaging system and method for glass | |
| US8351672B2 (en) | Machine imaging apparatus and method for detecting foreign materials | |
| EP1779096B1 (en) | Apparatus and method for checking of containers | |
| US5459330A (en) | Process and device for the inspection of glass | |
| US5774212A (en) | Method and apparatus for detecting and analyzing directionally reflective surface flaws | |
| CN106525873A (en) | Machine vision based full-automatic rotation printed product defect detection device | |
| WO1997046329A9 (en) | System and method for stress detection in a molded container | |
| US4678920A (en) | Machine vision method and apparatus | |
| CN110523657A (en) | A kind of instrument appearance delection device and its working method based on 3D scanning method | |
| CN211070921U (en) | Instrument appearance detection device based on 3D scanning method | |
| US20030035103A1 (en) | Method and device for inspecting transparent containers | |
| US7869021B2 (en) | Multiple surface inspection system and method | |
| CN113865830A (en) | Display screen defect detection method and system | |
| DE10297337T5 (en) | Automatic inspection apparatus and method for detecting anomalies in a 3-dimensional translucent object | |
| CN110412035A (en) | A kind of high reflecting material surface inspecting method and system | |
| KR20220134019A (en) | Systems and methods for imaging reflective objects | |
| JPH09264856A (en) | Article appearance inspection device | |
| CN109632810A (en) | Display panel crack detecting method and system | |
| JPH10160676A (en) | Rice grain inspection device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| EEER | Examination request | ||
| MKEX | Expiry |
Effective date: 20180123 |
|
| MKEX | Expiry |
Effective date: 20180123 |