WO2012027869A1

WO2012027869A1 - Online quality detection method for closure assembly

Info

Publication number: WO2012027869A1
Application number: PCT/CN2010/001735
Authority: WO
Inventors: 赵立品; 谢友顺; 梁劲; 杨艳梅; 黎万军; 冯科佳; 曹宇
Original assignee: 崇州君健塑胶有限公司
Priority date: 2010-08-31
Filing date: 2010-11-01
Publication date: 2012-03-08
Also published as: CN101975781A

Abstract

An online quality detection method for a closure assembly is provided, which uses a light source, cameras, image acquisition devices, a controller and a rejecting mechanism to detect the closure assembly. The method closely matches the production process of a closure assembling machine, enabling quality detection of both the inner and outer cover on the track of the closure assembling machine, and detection of the finished product after assembling the closure, the method rejects unqualified products by the controller; and can alarm and stop the closure assembling machine to work in the abnormal situation.

Description

Combined cover online quality inspection method

The present invention relates to a method of detection, and more particularly to an on-line quality inspection method for a combined lid production process.

Background technique

The combination cover is one of the production packaging materials necessary for the production of large infusions, and its demand in the infusion production process is very large. In the current domestic traditional combined cover production mode, in order to ensure the quality of the product, it is necessary to use a large number of workers to perform product quality inspection on the combined cover after the completion of the combined cover, and reject the non-conforming product in the combined cover. This method has many drawbacks, such as low production efficiency, high labor costs, unstable product quality, easy detection, and quality affected by the skill level of each tester. Summary of the invention

The invention aims to solve the problems of slow manual detection speed, low efficiency, unstable detection quality, introduction of secondary pollution in the detection process, and the like, and provides a combination of synchronous, online, non-contact, and automatic rejection of non-conforming products. Cover online quality detection method.

The invention is achieved by the following technical solutions:

A. Camera (Basler) imaging: The dedicated light source performs on-line quality inspection of the inner cover and the outer cover on the feeding track of the capping machine. By adding a transparent detection window on the track and cooperating with a dedicated light source, a plurality of industries Digital photography (Basler) imaging;

The dedicated light source illuminates the combined cover with a different wavelength of color light by using a top 360 degree uniform illumination with a full bottom imaging source;

B. Parallel image data acquisition, transmission and processing: After the sensor on the cover machine sends out the image signal, the three parts of the inner cover, the outer cover and the finished product start to collect images at the same time. According to the actual detection needs, the three stations can be used. Use up to 11 cameras (including zoom lenses) to capture images, and each camera (Basler) has its own independent channel for image transfer; when images are transferred from the capture card to the computer for processing, use DMA mode;

C. Elimination of non-conforming product: Through calculation and actual verification, the number of steps N1 (code) of the inner cover image collection point to the rejecting device, and the number of steps N2 (code) of the cover imaging point to the rejecting device, and the finished product imaging are obtained. Point to the number of steps N3 (code) at the culling device, each time the sensor is triggered, the corresponding Nl, N2, N3 count value minus 1, when the rejection point is reached, the three count values should be 0, then the inner, outer cover, finished product test results (pass or fail) to perform a logical "and" operation, get If the product is qualified, if it is a non-conforming product, the reject signal will be sent to control the solenoid valve action, and the non-conforming product will be removed into the designated non-conforming product box by compressed air.

The steps of the combined cover detecting method of the present invention may be:

1. Start the combined cover detection system computer (DELL T3500).

2. The cover machine is started by the control button. When the cover machine is running, the sensor (proximity switch) installed on the side of the cover machine evenly senses the indenter of the cover machine. When the cover of the cover machine is sensed, a high is generated. Level signal, and send this signal to each camera (Basler) through the camera (Basler) trigger line.

3. The camera (Basler) is connected to the computer (DELL T3500) via a universal 1394 frame grabber. Each of the three cameras (Basler) uses a 3-port image capture card to ensure that each camera has a separate transmission channel. After receiving the signal from the sensor, the camera captures the current image and transmits the image to the combo cover detection system computer via the 1394 data cable. The program in the combined cover detection system computer assigns separate threads to each camera, ensuring real-time image processing.

4. After the image enters the combined cover detection computer, the independent threads of the combined cover detection system computer program are separately processed to ensure parallel processing. In the thread, the image algorithm processing is performed, and the image algorithm processing uses the windows image processing module, and the specific processing algorithm may be:

a Image pre-processing: In order to remove unwanted data such as interference points and noise in the image, the erode (corrosion) and di late (expansion) algorithms in the image processing module are used once to effectively remove the interference points and noise of the image. Then, the image that needs to be processed is segmented, that is, the portion of the image that does not need to be detected is cropped. Use the Edge Detector algorithm to find the top, bottom, left and right borders of the image to be processed, so that the image can be accurately segmented.

b After the image is segmented, the image is binarized using the BackGround Correction algorithm in the image processing module. Then, the Close algorithm is called to process the interference points and noise that may occur in the binarization process to obtain a standard binarized image.

c After obtaining the standard binarized image, it is processed by the regional morphology and Part ic le Analys is processed in the image processing module to obtain impurity data information, including impurities, from the image. Information such as size, length and width, location.

d According to the program impurity detection standard, according to the area is larger than the number, the aspect ratio is less than the number to determine whether the image is qualified, and then the processing result is given.

e For the processing of the image acquired by the finished camera (Basler), using the Pater tern Ma tching algorithm, the acquired image is compared with the standard image to obtain a result value. If the value is less than a certain score, It is determined that the image is unqualified, and the processing result of whether the current combined cover is qualified is obtained, and the result is stored in the result queue corresponding to the camera (Basler).

5. Combined cover detection computer program The main process takes the processing result from the result queue corresponding to each camera, and performs a logical "AND" operation to obtain the final result of the product's eligibility. If the result is unqualified, pass the camera (Bas The trigger line sends a high level signal to the programmable logic controller.

6. Unqualified product rejection: After the rejection signal is transmitted to the programmable logic controller, the programmable logic controller controls to generate an output signal, which controls the solenoid valve to open for several tens of milliseconds, and the compressed air will be inferior during this time. Remove to the specified location.

7. The combined cover detection computer adopts the WINDOWS operating system, and the aforementioned program based on the WINDOWS operating system is composed of image acquisition, image processing and transmission command modules. The image acquisition module receives images captured by the camera and sent to the computer. The image processing module processes the acquired image and produces a result. The send command module sends this result to the programmable logic controller.

Due to the rapid development of optical imaging technology and computer processing technology, it provides a powerful guarantee for high-speed online quality detection based on optical imaging and computer image analysis technology. The present invention uses multiple digital signals in order to realize 360-degree detection of the outer side of the combined cover. At the same time, the surface quality of the camera is detected, and the detection result of each camera (Basler) is logically processed by the computer, and then the defective product is removed.

The invention is characterized in that it closely cooperates with the production process of the capping machine, realizes the quality inspection of the inner cap and the outer cap on the feeding track of the cap, and detects the quality of the finished product after the cap is completed, and finally realizes the automatic Elimination of non-conforming products, the integrated cover online quality inspection system has achieved breakthroughs in processing large amounts of concurrent data and a large number of graphics technologies, enabling the detection system to cooperate with the high-speed capping machine to complete the production, inspection and automatic sorting process of the combined cover. And monitor the system's operating status in real time during the production process. When an abnormal situation occurs, the system automatically alarms and stops, greatly improving the life of the capping machine. Production pass rate. In order to realize the parallel processing of a large amount of graphic data, the online detection system utilizes multi-core parallel technology, hyper-threading computing technology, real-time and efficient quality detection algorithm, high-speed programmable controller, etc., ensuring the completion of products in each group cover production cycle. Online quality inspection, and the detection system can automatically synchronize with the production speed of the combined cover, and ensure that the detection system can still work normally at maximum speed.

The invention has the following technical effects:

The invention relates to a contactless, real-time and online combined cover detecting system, which automatically realizes the quality inspection of the product and the automatic rejection of the non-conforming product in the production process of the capping machine, and can modify the product according to the quality requirements of the customer. Quality inspection conditions to meet the most demanding quality inspection requirements. Through the actual production use certificate, the quality inspection system can effectively detect the following various non-conformities:

> Bubbles, black spots, color points, scratches, burn marks, etc. on the inner cover.

> Bubbles, black spots, color points, scratches, smears, etc. on the cover.

Light cover, light inner cover, optical pad, etc.

> The finished product lacks inner cover, lack of rubber pad, severe crushing, and the combination is not in place.

DRAWINGS

Figure 1 Synthetic cover online quality detection schematic

Figure 2 Schematic diagram of the online quality inspection process of the combined cover

Figure 3 Combination cover dedicated imaging lighting system

Figure 4 Parallel image data acquisition, transmission, processing principle

Figure 5 Principle of rejecting nonconforming product

Figure 6 culling logical structure diagram

Figure Ί Impurity detection logic structure diagram

The invention is described in detail below with reference to the drawings and specific embodiments.

DETAILED DESCRIPTION 1

1 and 2 are schematic diagrams showing the principle of the online quality detecting system of the combined cover of the present invention, including on-line detection of the inner cover, on-line detection of the outer cover, online detection of the finished product, and online rejection mechanism. In the specific implementation, the online detection system performs on-line quality inspection of the inner cover and the outer cover on the feeding track of the capping machine, by adding a transparent detection window on the track, and matching a dedicated machine vision light source. Provides the best imaging environment for multiple industrial digital cameras.

The dedicated light source used in the combined cover imaging is shown in Figure 3 (combined cover-specific imaging lighting system). The material of the combined cover determines that it is a translucent imaging condition, using a top 360 degree uniform illumination with a full bottom. The imaging light source illuminates the combined cover with different wavelengths of colored light. After the different colored lights are absorbed, reflected and superimposed on the combined cover, the impurities are highlighted and sharpened, and some interference images inherent in the combined cover are weakened (such as The shadows and thickness variations at the corners are shaded, etc.), providing the best image for later computer analysis.

In the process of detection, in order to realize parallel processing of data of multiple cameras (Basler), special design and implementation are implemented in the data transmission channel, parallel computing processing unit, image quality detection software, etc., to ensure the shortest specified Complete the entire quality inspection process during the cycle. The parallel processing implementation is as shown in Figure 4 (parallel image data acquisition, transmission, processing principle). After the sensor on the cover machine sends out the image signal, the three parts of the inner cover, the outer cover and the finished product start to collect images at the same time. 3 stations can use up to 11 cameras (including zoom lenses) to capture images according to the actual detection needs, and each camera (Basler) has a separate channel for image transmission, and the image is collected. When the card passes the data to the computer for processing, the DMA method is used to reduce the intervention of the real-time controller CPU on the data transfer, and more CPU computing power is used for the quality detection algorithm and the non-conforming product rejection algorithm.

In order to achieve full quality inspection of the entire composite cover, each camera (Basler) is responsible for completing the product quality inspection within its monitored range, and when the final determination of the product is acceptable, the detection results of each camera (Basler) are compared and operated. , to determine whether the product is qualified. After the real-time computer obtains the image taken by each camera, the image is pre-processed to remove unnecessary data such as interference points and noises in the image, and at the same time, the image portion to be detected is positioned and segmented. Out of the image that needs to be processed. Then, according to the gray value characteristics of the image, the image is binarized, and the impurity data information is obtained from the image through the analysis process of the region morphology and the like, and finally determined according to the geometric features of the impurity and the set impurity detection standard. The camera detects whether the part is qualified. After the image data of each camera is processed, the detection result is sent to the product rejection part for automatic rejection of the defective product.

In order to finally realize the rejection of the non-conforming products in the composite cover product, a set of non-conforming product elimination logic algorithm and rejection mechanism was designed. Due to the detection position of the inner and outer covers and the detection position and final result of the finished product The positions of the rejects are different, and the number of steps of the inner cover, the outer cover, and the finished product to the reject position are different. The elimination logic algorithm ensures that the detection results of each camera can be accurately reflected to the final judgment product. Whether the qualified control signal is passed on, and the signal is transmitted to the programmable logic controller, and the mechanical execution part is controlled to eliminate the defective product.

DETAILED DESCRIPTION 2

The steps of the combined cover detection method are as follows:

1. Start the combined cover detection system computer (DELL T3500), and the combined cover detection system program starts running.

3. The camera (Basler) is connected to the computer (DELL T3500) via a universal frame grabber. Each of the three cameras (Basler) uses a 3-port image capture card to ensure that each camera (Basler) has a separate transmission channel. After receiving the signal from the sensor, the camera acquires the current image and transmits the image to the combined cover detection system program via the Basler 1394 data cable. The combined cover detection system program assigns a separate thread to each camera (Basler), ensuring real-time image processing.

4. After the image enters the combined cover detection system program, it is processed separately by the aforementioned independent threads to ensure parallel processing. In the thread, the image algorithm processing is performed, and the image algorithm processing uses the windows image processing module, and the specific processing algorithm can be:

a Image pre-processing: In order to remove unwanted data such as interference points and noise in the image, the erode (corrosion) and di late (expansion) algorithms in the image processing module are used once to effectively remove the interference points and noise of the image. Then, the image that needs to be processed is segmented, that is, the portion of the image that does not need to be detected is cropped. Use the Edge Detector algorithm to find the top, bottom, left, and right edges of the image to be processed, so that the image can be accurately segmented.

b After the image is segmented, the image is binarized using the BackGround Correction algorithm in the image processing module. Then, the Close algorithm is called to process the interference points and noise that may be generated during the binarization process to obtain a standard binarized image. c After obtaining the standard binarized image, it is processed by the region morphology and the Paricle Analys is algorithm in the image processing module to obtain the impurity data information from the image, including the size, length and width of the impurity, and the location. And other information.

e For the processing of the finished image of the Basler, use the Pat tern Ma tching algorithm to compare the acquired image with the standard image to obtain a result value. If the value is less than a certain value The score, that is, the image is judged to be unqualified, and the result of the processing of the current combined cover is obtained, and the result is stored in the result queue corresponding to the shot.

5. The combined cover detection system program main process takes the processing result from the result queue corresponding to each camera, and performs a logical "AND" operation to obtain the final result of the product's eligibility. If the result is unqualified, pass the camera (Bas The trigger line sends a high level signal to the programmable logic controller.

7. The above is a complete monitoring process. At present, the system can achieve such monitoring every second.

Claims

Rights request

A method for detecting online quality of a combined cover, characterized in that the method comprises the following steps:

A. Camera imaging: On-line quality inspection of the inner cover and the outer cover on the feeding track of the capping machine, by adding a transparent detection window on the track, and using a dedicated light source, imaging by multiple industrial digital cameras;

B. Parallel image data acquisition, transmission and processing: After the sensor on the cover machine sends out the image signal, the three parts of the inner cover, the outer cover and the finished product start to collect images at the same time. According to the actual detection needs, the three stations can be used. Use up to 11 cameras to capture images, and each camera has its own independent channel for image transfer; when the image is transferred from the capture card to the computer for processing, the DMA method is used;

C. Elimination of non-conforming product: Through calculation and actual verification, obtain the step value C1 of the inner cover image collection point to the rejecting device, and the step value C2 of the cover imaging point to the rejecting device, and the finished imaging point to the rejecting device. Step value C3, each time the sensor is triggered, the corresponding Cl, C2, C3 count value is decremented by 1. When the reject point is reached, the three count values should be 0. At this time, the detection results of the inner and outer covers and the finished product are executed. The series "and" operation, to obtain the signal of whether the product is qualified, if it is a non-conforming product, the rejection signal is sent to control the solenoid valve action, and the non-conforming product is rejected by the compressed air to the designated non-conforming product phase.

2. The combined cover online quality detecting method according to claim 1, wherein the dedicated light source of the method refers to the use of different wavelengths of color light to the combined cover by using the top 360 degree uniform illumination and the full bottom imaging light source. illumination.

3, a combined cover online quality detection method, characterized in that the method comprises the following steps: 启动. Start the combined cover detection system computer;

Β. The cover machine is started by the control button. When the cover machine is running, the sensor installed on the side of the cover machine evenly senses the head of the cover machine. When the cover of the cover machine is sensed, a high level signal is generated. And transmit this signal to each camera through the camera trigger line;

C. The camera is connected to the computer through a universal 1394 image capture card; each of the three cameras uses a 3-port image capture card to ensure that each camera has an independent transmission channel; after receiving the signal from the sensor, the camera captures the current image. And take this image through the 1394 data The line is transmitted to the combined cover detection system computer; the combined cover detection system computer program allocates independent threads to each camera to ensure real-time image processing;

D. After the image enters the combined cover detection system computer, it is processed by the independent thread separately to ensure parallel processing; in the thread, image algorithm processing is performed, and the image algorithm processing uses windows image processing module, and the specific processing algorithm is:

a image pre-processing: In order to remove the useless data such as interference points and noises in the image, the erode and di late algorithms in the image processing module are used once to effectively remove the interference points and noise of the image; then, the part to be processed is segmented. The image, that is, the part of the image that is not to be detected is cut out; using the Edge Detector algorithm, the upper and lower left and right borders of the image to be processed are found, so that the image can be accurately segmented;

After the image segmentation is completed, the image is binarized using the BackGround Correc t ion algorithm in the image processing module. Then, the Close algorithm is called to process the interference points and noises that may be generated during the binarization process, and the standard 2 is obtained. Valued image

c After obtaining the standard binarized image, it is processed by the region morphology and Part ic le Analys is algorithm in the image processing module, and the impurity data information is obtained from the image, including the size, length and width, location of the impurity, and the like;

d According to the program impurity detection standard, according to the area is larger than the number, the aspect ratio is less than the number to determine whether the image is qualified, and then the processing result is given;

e For the processing of the acquired image of the finished camera, using the Pat tern Matching algorithm, the image to be acquired is compared with the standard image, and a result value is obtained. If the value is less than a certain score, the image is judged to be unqualified, and the current result is obtained. The result of processing whether the combined cover is qualified, and storing the result in the result queue corresponding to the camera;

E. Combined cover detection system The main program of the computer program takes the processing result from the result queue corresponding to each camera, and performs a logical AND operation on the result to obtain the final result of whether the product is qualified. If the result is unqualified, it is triggered by the camera. The line sends a high level signal to the programmable logic controller;

F. reject of non-conforming product: After the rejection signal is transmitted to the programmable logic controller, the programmable logic controller controls to generate an output signal, which controls the solenoid valve to open for several tens of milliseconds, and the compressed air will be non-conforming during this time. Remove to the specified location.