CN103258331A - Position determination method of sheltered dynamic target - Google Patents
Position determination method of sheltered dynamic target Download PDFInfo
- Publication number
- CN103258331A CN103258331A CN2012104317168A CN201210431716A CN103258331A CN 103258331 A CN103258331 A CN 103258331A CN 2012104317168 A CN2012104317168 A CN 2012104317168A CN 201210431716 A CN201210431716 A CN 201210431716A CN 103258331 A CN103258331 A CN 103258331A
- Authority
- CN
- China
- Prior art keywords
- dynamic object
- shelter
- positional information
- blocking
- dynamic target
- 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.)
- Pending
Links
Images
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a position determination method of a sheltered dynamic target. The position determination method of the sheltered dynamic target includes the following steps: (1) judging whether the dynamic target is sheltered or not, (2) judging whether a shelter is static or moving, (3) describing the motion track of the dynamic target according to the acquired positional information of the dynamic target, estimating the average movement speed of the dynamic target, setting the position of the dynamic target at the moment when the dynamic target disappears as a sheltered position and determining the sheltered position when the shelter is static, (4) determining the sheltered position through the above method when the shelter is moved and is an instant shelter, and (5), calculating the sheltered position of the dynamic target when the shelter moves according to the positional relation between the sheltered position of the dynamic target and the movable shelter when the shelter occurs if the shelter is a following shelter. By means of the position determination method of the sheltered dynamic target, the sheltered position of the dynamic target can be calculated unceasingly when the dynamic target is sheltered, and the position information of the dynamic target can be supplied to a recognition system unceasingly.
Description
Technical field
The present invention relates to identification and the camera calibration technology of image in the artificial intelligence field, be specifically related to carry out image when location target position calculation method of devising a stratagem really when being blocked with video camera.
Background technology
As everyone knows, the mankind are by eyes, brain obtains, and handle and understand visual information, and image vision information is one of human main source from the objective event acquired information.Video image is combined by the still image of a series of contacts in fact exactly, and he is the description vivid to objective things, is a kind of directly perceived and concrete information representation form.Characteristics such as along with the develop rapidly of computer network, communication and microelectric technique, video monitoring is directly perceived with it, and the convenient and information content is abundant are subject to people's attention day by day and study.
As: the dynamic object capturing technology has a wide range of applications at artificial intelligence field, and utilizing video camera that the picture dynamic object is caught and reduce its physical location in the space is the key that engineering can be seen.
To some simple situations, computing machine can be realized the location of dynamic object, but situation about being blocked for target, and because picture can't show the target that is blocked, thereby location algorithm may lose efficacy.If system requirements is known the positional information of this target, will produce unpredicted problem so.
The technology for eliminating that blocks in the past all is in the operation of pixel scale, uses method of difference and filtration method mostly, though the good results are evident, belongs to technical field.It is the problem that this area is needed solution badly that a kind of occlusion handling method that belongs to artificial intelligence field is provided thus.
Summary of the invention
The present invention is directed in the existing artificial intelligence field and when dynamic object is identified, blocking the existing problem of processing scheme, position calculating method and the dynamic object that a kind of visual theory based on the people is provided is blocked under the situation, can under target is blocked condition, carry out logic and judge, can well be applicable to artificial intelligence field.
In order to achieve the above object, the present invention adopts following technical scheme:
A kind of dynamic object is in the location determining method that has under the situation of blocking, and this method comprises the steps:
(1) when the dynamic object positional information disappears, dynamic object positional information harvester detects the positional information of dynamic object in real time, in threshold time, if obtain the positional information of dynamic object again, and the positional information when disappearing with dynamic object is different, is judged as dynamic object and is blocked; If do not obtain the positional information of dynamic object again, it is static then to be judged as dynamic object;
(2) after definite dynamic object is blocked, the state of judging shelter is static still mobile, if shelter is the static step (3) that changes over to; If shelter is the mobile step (4) that then changes over to;
(3) if shelter is stationary state, depict the movement locus of dynamic object according to the positional information that obtains dynamic object, and estimate its average translational speed, disappear from dynamic object and to begin set positions with dynamic object that time for blocking the position, and by the dynamic object translational speed, time, the track geometric relationship determines to block the position;
(4) if shelter is mobile status, the shelter of judging mobile status according to the length of blocking the duration to dynamic object be instantaneous block or follow block; Very short as if blocking the duration, then be judged as instantaneous blocking, and change step (5) over to; If block longer duration, then be judged as to follow and block, and change step (6) over to;
(5) if mobile shelter is instantaneous blocking to dynamic object, then adopt the mode of step (3) to determine to block the position;
(6) if mobile shelter blocks for following dynamic object, to block the position and concern with respect to the position of mobile shelter according to blocking dynamic object when taking place, dynamic object blocks the position when calculating shelter and moving.
In preferred embodiment of the present invention, determine the positional information of dynamic object in the above-mentioned steps in real time by formula (1):
X wherein, Y is the pixel coordinate of object in the image frame, x
w, y
wBe object at the actual coordinate on ground, L is determined with reference to coefficient by pre-scaling method.
According to the present invention that such scheme forms, its dynamic object based on people's visual theory position calculating method under the situation that is blocked namely carries out logic and judges under target is blocked condition.The present invention can be when dynamic object be blocked, and continues to calculate the position of blocking of dynamic object, can continue as the positional information that recognition system continues to provide dynamic object, and effectively solve picture and can't show the target that is blocked, and problem that can't location algorithm.
Description of drawings
Further specify the present invention below in conjunction with the drawings and specific embodiments.
Fig. 1 is process flow diagram of the invention process.
Embodiment
For technological means, creation characteristic that the present invention is realized, reach purpose and effect is easy to understand, below in conjunction with concrete diagram, further set forth the present invention.
People's vision is to realize after target disappears from the visual field the judgement of its position by brain to the continuous memory of target location, as long as this principle shows track and the mobile speed that allows computing machine can note dynamic object, just can in the short time that target disappears, estimate its position.
Based on above-mentioned principle, concrete enforcement of the present invention is as follows:
In the 1st step, judge whether to block
Because the positional information of dynamic object can suddenly disappear when blocking generation, therefore when the dynamic object positional information disappears, system is judged as that it is static or be blocked, but block with static unique difference and be that the position of target is different after a period of time, if there is a segment distance when target position information occurs again and the position of its disappearance, system just is judged as target and has been blocked, otherwise is exactly static.
Thus, the specific implementation of this step is: when the dynamic object positional information disappears, dynamic object positional information harvester will continue to detect in real time the positional information of dynamic object, and the time of record dynamic object positional information disappearance, if in certain threshold value (this threshold time can come according to the actual requirements calmly) time, if obtain the positional information of dynamic object again, positional information when the positional information of the dynamic object that obtains is again disappeared with the dynamic object positional information compares: inequality as if both, and be judged as dynamic object and be blocked; If both are identical, it is static then to be judged as dynamic object, and static processing will be done to dynamic object by system this moment.
The 2nd step, the state of judgement shelter
Because it is a variety of that situation about being blocked has, can be divided into shelter according to the state classification of shelter is stationary object, and shelter is mobile object.According to be blocked thing from block beginning takes place may be static, by the motion of moving or shift gears of original mode, different thus mode classifications is very big to the position judgment influence of the object that is blocked, so the present invention has proposed effective disposal route respectively to various situations.
This step detects in real time by the positional information to shelter after definite dynamic object is blocked, and carries out front and back relatively, and the state of judging shelter is static still mobile, if shelter is that the static step (3) that changes over to is handled; If shelter is handled for the mobile step (4) that then changes over to.
In the 3rd step, statistics dynamic object method of operation information determines that dynamic object blocks the position
Be the situation of stationary object for shelter, according to blocking the movement locus that the positional information that obtains dynamic object before taking place is depicted dynamic object, and estimate its average translational speed with this; Disappearing from target begins set positions with dynamic object that time for blocking the position, wherein blocks the position according to the translational speed of dynamic object, blocks the corresponding track geometric relationship of time and movement locus accordingly and comprehensively determines.
In order further to improve the degree of accuracy of blocking position calculation, this step is further judged the type of blocking.As shown in the figure, in step 31, according to the length of blocking the duration judge static shelter to dynamic object be instantaneous block or follow block; Very short as if blocking the duration, then be judged as instantaneous blocking, and change step (32) over to; If not, then change step (33) over to.
Export corresponding inertial coordinate in the step (32), and determine the position of dynamic object with this.
Step (33) judges that this dynamic object constantly plays from being blocked that to walk out shelter be static in finishing during this period of time.
In the 4th step, at mobile shelter, judge the type that it blocks dynamic object
Be the situation of moving object for shelter, at first the length of the duration of dynamic object being blocked according to mobile shelter judge mobile shelter to dynamic object be instantaneous block or follow block; Very short as if blocking the duration, then be judged as instantaneous blocking, the position that can adopt the mode in the above-mentioned steps (3) to determine to calculate dynamic object; If block longer duration, then be judged as to follow and block, and change step (5) over to and handle.
In the 5th step, determine dynamic destination locations
Dynamic object blocks the position with respect to the position relation of mobile shelter when taking place according to blocking in this step, utilizes corresponding geometric relationship to form corresponding position coordinates, and when calculating shelter and move with this dynamic object block the position.
At continuing to block the problem that some situation can't be judged in the situation, have a mind to avoid sight line such as target and move in the present invention, it is static perhaps to avoid sight line, and these two kinds of situations have no idea to judge continuing to block down.The present invention is by arranging the situation that the normal route that can avoid dynamic object in the video camera is blocked fully by object in the middle of reality, therefore continued to block in this case by stationary object in target, the present invention sets this dynamic object and plays from being blocked that to walk out shelter be static in finishing during this period of time constantly.
Such scheme needs to determine by the image picture in real time the physical location of dynamic object in the process of implementing, the recognition methods of this following image physical location in the present invention, and this method is carried out based on formula 1:
X wherein, Y is the pixel coordinate of object in the image frame, x
w, y
wBe object at the actual coordinate on ground, L is determined with reference to coefficient by pre-scaling method.
Be exactly a kind of system antidote after parameters such as camera position change for pre-scaling method.In the present invention, by formula (1) is carried out conversion, obtain two about the system of linear equations of coefficient L:
Because this system of equations comprises 8 unknown numbers, therefore want to obtain fully the equation that coefficient just needs 8 correspondences.Notice that each ground coordinate and corresponding picture coordinate have the relation of (1) formula, so each just provides 2 equations to such point, the present invention utilizes 4 pre-calibration points to come the coefficient of each L of initialization, after obtaining 8 systems of linear equations, use Gauss _ remove when the elimination method approximately coefficient of each L.
The L coefficient that the above-mentioned pre-scaling method of recycling obtains has obtained by formula (2)
When setting up, ground coordinate system got well the physical location that the coordinate that utilizes later formula (3) formula representative has just provided ground point.
Certainly system of linear equations is not when well-determined solution to be arranged, in the time of more special in the presence of the position of four points, such as at least three points point-blank, and unique 8 above-mentioned coefficients of determining of just having no idea at this time just require to demarcate again four points.
After determining, each L coefficient gets final product about the relational expression of picture pixel coordinate by the anti-ground coordinate that solves of (1) formula again.
This conversion purpose only is just to calculate it in the actual position coordinate on ground by the pixel coordinate of picture, and is indifferent to inside and the external parameter of video camera, can satisfy corresponding application scenarios.
Recognition methods based on above-mentioned image physical location, an if some P (2 in the given camera picture, 2), it is (2,2) at the pixel coordinate of picture so, after transfer process of the present invention, obtained ground coordinate p (10,10), this p is exactly the ground physical location that the camera picture meta is set to the some P representative of (2,2) so.
By such scheme, the visual theory that the present invention is based on the people determines to calculate be blocked position under the situation of dynamic object, can the logic of carrying out under target is blocked condition judge, is applicable to artificial intelligence field fully.
More than show and described ultimate principle of the present invention, principal character and advantage of the present invention.The technician of the industry should understand; the present invention is not restricted to the described embodiments; that describes in above-described embodiment and the instructions just illustrates principle of the present invention; without departing from the spirit and scope of the present invention; the present invention also has various changes and modifications, and these changes and improvements all fall in the claimed scope of the invention.The claimed scope of the present invention is defined by appending claims and equivalent thereof.
Claims (2)
1. a dynamic object is characterized in that in the location determining method that has under the situation of blocking, and described method comprises the steps:
(1) when the dynamic object positional information disappears, dynamic object positional information harvester detects the positional information of dynamic object in real time, in threshold time, if obtain the positional information of dynamic object again, and the positional information when disappearing with dynamic object is different, is judged as dynamic object and is blocked; If do not obtain the positional information of dynamic object again, it is static then to be judged as dynamic object;
(2) after definite dynamic object is blocked, the state of judging shelter is static still mobile, if shelter is the static step (3) that changes over to; If shelter is the mobile step (4) that then changes over to;
(3) if shelter is stationary state, depict the movement locus of dynamic object according to the positional information that obtains dynamic object, and estimate its average translational speed, disappear from dynamic object and to begin set positions with dynamic object that time for blocking the position, and by the dynamic object translational speed, time, the track geometric relationship determines to block the position;
(4) if shelter is mobile status, the shelter of judging mobile status according to the length of blocking the duration to dynamic object be instantaneous block or follow block; Very short as if blocking the duration, then be judged as instantaneous blocking, and change step (5) over to; If block longer duration, then be judged as to follow and block, and change step (6) over to;
(5) if mobile shelter is instantaneous blocking to dynamic object, then adopt the mode of step (3) to determine to block the position;
(6) if mobile shelter blocks for following dynamic object, to block the position and concern with respect to the position of mobile shelter according to blocking dynamic object when taking place, dynamic object blocks the position when calculating shelter and moving.
2. a kind of dynamic object according to claim 1 is characterized in that in the location determining method that has under the situation of blocking, and determines the positional information of dynamic object in the described step in real time by formula (1):
X wherein, Y is the pixel coordinate of object in the image frame, x
w, y
wBe object at the actual coordinate on ground, L is determined with reference to coefficient by pre-scaling method.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012104317168A CN103258331A (en) | 2012-11-01 | 2012-11-01 | Position determination method of sheltered dynamic target |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2012104317168A CN103258331A (en) | 2012-11-01 | 2012-11-01 | Position determination method of sheltered dynamic target |
Publications (1)
Publication Number | Publication Date |
---|---|
CN103258331A true CN103258331A (en) | 2013-08-21 |
Family
ID=48962224
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2012104317168A Pending CN103258331A (en) | 2012-11-01 | 2012-11-01 | Position determination method of sheltered dynamic target |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN103258331A (en) |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114529858A (en) * | 2022-04-21 | 2022-05-24 | 浙江大华技术股份有限公司 | Vehicle state recognition method, electronic device, and computer-readable storage medium |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2000011182A (en) * | 1997-10-03 | 2000-01-14 | Nippon Telegr & Teleph Corp <Ntt> | Time-sequential image characteristics extraction method and device therefor and record medium for recording the same method |
US6236737B1 (en) * | 1997-03-26 | 2001-05-22 | Dalhousie University | Dynamic target addressing system |
US20040156530A1 (en) * | 2003-02-10 | 2004-08-12 | Tomas Brodsky | Linking tracked objects that undergo temporary occlusion |
CN1968352A (en) * | 2005-11-17 | 2007-05-23 | 中国科学院半导体研究所 | High-speed target tracking method and its circuit system |
CN102737386A (en) * | 2012-05-15 | 2012-10-17 | 北京硅盾安全技术有限公司 | Moving target anti-fusion shielding tracking algorithm |
-
2012
- 2012-11-01 CN CN2012104317168A patent/CN103258331A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6236737B1 (en) * | 1997-03-26 | 2001-05-22 | Dalhousie University | Dynamic target addressing system |
JP2000011182A (en) * | 1997-10-03 | 2000-01-14 | Nippon Telegr & Teleph Corp <Ntt> | Time-sequential image characteristics extraction method and device therefor and record medium for recording the same method |
US20040156530A1 (en) * | 2003-02-10 | 2004-08-12 | Tomas Brodsky | Linking tracked objects that undergo temporary occlusion |
CN1968352A (en) * | 2005-11-17 | 2007-05-23 | 中国科学院半导体研究所 | High-speed target tracking method and its circuit system |
CN102737386A (en) * | 2012-05-15 | 2012-10-17 | 北京硅盾安全技术有限公司 | Moving target anti-fusion shielding tracking algorithm |
Non-Patent Citations (1)
Title |
---|
郑世友等: "基于仿射不变量的运动目标遮挡检测和跟踪", 《信号处理》, vol. 22, no. 3, 30 June 2006 (2006-06-30), pages 434 - 439 * |
Cited By (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN114529858A (en) * | 2022-04-21 | 2022-05-24 | 浙江大华技术股份有限公司 | Vehicle state recognition method, electronic device, and computer-readable storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN103077539B (en) | Motion target tracking method under a kind of complex background and obstruction conditions | |
CN105447458B (en) | A kind of large-scale crowd video analytic system and method | |
CN102006461B (en) | Joint tracking detection system for cameras | |
CN108198206A (en) | The multi-object tracking method combined based on multiple features combining and Camshift algorithms | |
CN102930248A (en) | Crowd abnormal behavior detection method based on machine learning | |
CN104700430A (en) | Method for detecting movement of airborne displays | |
CN105046719B (en) | A kind of video frequency monitoring method and system | |
CN105184812B (en) | A kind of pedestrian based on target following hovers detection method | |
CN105678803A (en) | Video monitoring target detection method based on W4 algorithm and frame difference | |
CN103530874A (en) | People flow counting method on basis of Kinect | |
JP2007142527A (en) | Mobile body supervisory system, mobile body feature quantity calculation apparatus, and inter-camera mobile body collation tracing apparatus | |
CN103258332A (en) | Moving object detection method resisting illumination variation | |
CN102855465B (en) | A kind of tracking of mobile object | |
CN102880444A (en) | Fighting detecting method based on stereoscopic vision motion field analysis | |
Berdnikov et al. | Real-time depth map occlusion filling and scene background restoration for projected-pattern-based depth cameras | |
CN112561951B (en) | Motion and brightness detection method based on frame difference absolute error and SAD | |
CN101483001A (en) | Video-based intrusion detection method, detection apparatus and chip | |
CN106251363A (en) | A kind of wisdom gold eyeball identification artificial abortion's demographic method and device | |
CN103458261A (en) | Video scene variation detection method based on stereoscopic vision | |
CN103425960B (en) | Fast moving objects method for detecting in a kind of video | |
CN103927519A (en) | Real-time face detection and filtration method | |
CN107295296A (en) | A kind of selectively storage and restoration methods and system of monitor video | |
CN107509059A (en) | Camera lens occlusion detection method | |
CN114037087B (en) | Model training method and device, depth prediction method and device, equipment and medium | |
Liu et al. | Abnormal crowd behavior detection based on optical flow and dynamic threshold |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20170901 |
|
AD01 | Patent right deemed abandoned |