CN106530325A - Multi-target visual detection and tracking method - Google Patents
Multi-target visual detection and tracking method Download PDFInfo
- Publication number
- CN106530325A CN106530325A CN201610946656.1A CN201610946656A CN106530325A CN 106530325 A CN106530325 A CN 106530325A CN 201610946656 A CN201610946656 A CN 201610946656A CN 106530325 A CN106530325 A CN 106530325A
- Authority
- CN
- China
- Prior art keywords
- target
- tracking
- model
- detection
- image
- 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
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Abstract
The invention discloses a multi-target visual detection and tracking method. The method comprises the steps that 1 a motion target is detected, and the location, the size and the binary mask image of the target are monitored; 2 global features are extracted, and the target is described through a color histogram, a gradient direction histogram and an LBP texture appearance model; 3 target inter-frame correlation matching is carried out; and 4 a target tracking model is used to acquire the motion trajectory of the target. According to the invention, a number of targets are simultaneously detected through the method based on a motion detection target; the detection cost is reduced; target correlation tracking based on the color histogram and other global features is used; if the target is occluded, model matching is carried out through the color histogram which occludes the target; and the detection accuracy is improved.
Description
Technical field
The invention belongs to computation vision detection technique field, more particularly to a kind of multiple target vision-based detection and track side
Method.
Background technology
Vision is the important means in mankind's observation and the cognitive world.There are about in the information that the mankind are obtained from outside 75% from
In visual system.Computer vision is as an independent subject if being to teach this from Massachusetts Institute Technology Man
Start for the work of laying a foundation done by people, then its own Jing has the history of nearly 40 years.Due to computer vision potential application very
Extensively, involved subject knowledge is extremely various, and the problem of research is rich in challenge again, therefore it is always in Computer Subject
One popular research.In recent years, with the development of image processing techniquess and artificial intelligence, computer vision technique is also constantly
Progressive, its application is also constantly extending.
Visual target tracking is an important task in computer vision, and so-called visual target tracking is exactly to figure
As the moving target in sequence is detected, recognized and tracked, the kinestate of target is obtained：Position, size, speed and target
Movement locus etc..By the analysis of the state to moving target and track, some more high-rise behavior analysiss can be carried out,
Common application includes：Intelligent security-protecting and monitoring, intelligent transportation, man-machine interaction and intelligent navigation etc..Visual target tracking is in intelligence
It is mainly used in safety monitoring and the viewing areas such as security department, public arena, private residence is exercised supervision, to questionable conduct
Carry out detection being carried out with abnormal conditions and report is shot a glance at automatically：The virgin statistics of traffic flow is mainly used in intelligent transportation and vehicle is different
Reason condition detection etc.；Identification of the behaviors such as gesture, attitude and action etc. is mainly used in terms of man-machine interaction：Intelligent navigation side
The robotic tracking visual field middle finger that is mainly used in face sets the goal.
The content of the invention
It is an object of the invention to provide a kind of multiple target vision-based detection and tracking, by based on motion detection target
Method, simultaneously multiple targets can be detected, be tracked using the target association based on global characteristics such as color histograms,
Occur, under circumstance of occlusion, to carry out Model Matching by the color histogram of shelter target, improve the accuracy of detection in target.
The present invention is achieved by the following technical solutions：
The present invention is a kind of multiple target vision-based detection and tracking, is comprised the steps：
Step one, moving object detection：The position of monitoring objective, target sizes and target two-value mask image；
Step 2, global characteristics are extracted：By the display model of color histogram, gradient orientation histogram and LBP textures
Target is described；
Step 3, the matching of target intra-frame trunk：By finding a closest detection target to each tracking target,
When this distance is in the range of given threshold value, then it represents that be matching between the tracking target and detection target；
Step 4, target following model；The movement locus of target are obtained by target following model.
Preferably, the moving object detection is concretely comprised the following steps：
Current frame image by Gaussian smoothing denoising, is obtained smoothed image by a；
B carries out the background modeling of image by time average background model and/or mixture Gaussian background model, by modeling
Obtain prospect bianry image and background image；
C shadow removals, the shade of prospect bianry image is removed；
D mass detections, obtain position and the size of agglomerate.
Preferably, the method for the shadow removal is using the method based on model or the method based on attribute；Based on model
Method be to set up shadow model by using scene, target and illumination, rib, angle to three-dimensional body, line are matched；It is based on
The method of attribute is indicating shadow region using the geometrical feature of shade, brightness, color information.
Preferably, the mass detection uses the shadow detection method of Rita Cucchiara.
Preferably, in the step 4, target following model adopts Kalman filter.
The invention has the advantages that：
The present invention is by based on motion detection mesh calibration method, can detect to multiple targets, reduce inspection simultaneously
Cost is surveyed, is tracked using the target association based on global characteristics such as color histograms, occurred under circumstance of occlusion in target, by hiding
The color histogram of gear target carries out Model Matching, improves the accuracy of detection.
Certainly, the arbitrary product for implementing the present invention is it is not absolutely required to while reaching all the above advantage.
Description of the drawings
In order to be illustrated more clearly that the technical scheme of the embodiment of the present invention, use required for describing to embodiment below
Accompanying drawing is briefly described, it should be apparent that, drawings in the following description are only some embodiments of the present invention, for ability
For the those of ordinary skill of domain, on the premise of not paying creative work, can be attached to obtain others according to these accompanying drawings
Figure.
Fig. 1 is a kind of multiple target vision-based detection and tracking flow chart of the present invention；
Flow charts of the Fig. 2 for moving object detection.
Specific embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is only a part of embodiment of the invention, rather than the embodiment of whole.It is based on
Embodiment in the present invention, it is all other that those of ordinary skill in the art are obtained under the premise of creative work is not made
Embodiment, belongs to the scope of protection of the invention.
Refer to shown in Fig. 1, the present invention is a kind of multiple target vision-based detection and tracking, is comprised the steps：
Step one, moving object detection：The position of monitoring objective, target sizes and target two-value mask image；
Step 2, global characteristics are extracted：By the display model of color histogram, gradient orientation histogram and LBP textures
Target is described；
Step 3, the matching of target intra-frame trunk：By finding a closest detection target to each tracking target,
When this distance is in the range of given threshold value, then it represents that be matching between the tracking target and detection target；
Step 4, target following model；The movement locus of target are obtained by target following model.
Wherein as shown in Fig. 2 moving object detection is concretely comprised the following steps：
Current frame image by Gaussian smoothing denoising, is obtained smoothed image by a；
B carries out the background modeling of image by time average background model and/or mixture Gaussian background model, by modeling
Obtain prospect bianry image and background image；
C shadow removals, the shade of prospect bianry image is removed；
D mass detections, obtain position and the size of agglomerate.
Wherein, the method for shadow removal is using the method based on model or the method based on attribute；Method based on model
It is to set up shadow model by using scene, target and illumination, rib, angle to three-dimensional body, line are matched；Based on attribute
Method is indicating shadow region using the geometrical feature of shade, brightness, color information.
Wherein, mass detection uses the shadow detection method of Rita Cucchiara.
Wherein, in step 4, target following model adopts Kalman filter.
It should be noted that in said system embodiment, included unit simply carries out drawing according to function logic
Point, but above-mentioned division is not limited to, as long as corresponding function can be realized；In addition, each functional unit is concrete
Title is also only to facilitate mutually differentiation, is not limited to protection scope of the present invention.
In addition, one of ordinary skill in the art will appreciate that realizing all or part of step in the various embodiments described above method
Program be can be by instruct the hardware of correlation to complete, corresponding program can be stored in embodied on computer readable storage and be situated between
In matter, described storage medium, such as ROM/RAM, disk or CD etc..
Present invention disclosed above preferred embodiment is only intended to help and illustrates the present invention.Preferred embodiment is not detailed
All of details is described, it is only described specific embodiment also not limit the invention.Obviously, the content according to this specification,
Can make many modifications and variations.These embodiments are chosen and specifically described to this specification, is to preferably explain the present invention
Principle and practical application so that skilled artisan can be best understood by and utilize the present invention.The present invention is only
Limited by claims and its four corner and equivalent.
Claims (5)
1. a kind of multiple target vision-based detection and tracking, it is characterised in that comprise the steps：
Step one, moving object detection：The position of monitoring objective, target sizes and target two-value mask image；
Step 2, global characteristics are extracted：By the display model of color histogram, gradient orientation histogram and LBP textures to mesh
Mark is described；
Step 3, the matching of target intra-frame trunk：By finding a closest detection target to each tracking target, when this
Individual distance is in the range of given threshold value, then it represents that be matching between the tracking target and detection target；
Step 4, target following model；The movement locus of target are obtained by target following model.
2. a kind of multiple target vision-based detection according to claim 1 and tracking, it is characterised in that the moving target
That what is detected concretely comprises the following steps：
Current frame image by Gaussian smoothing denoising, is obtained smoothed image by a；
B carries out the background modeling of image by time average background model and/or mixture Gaussian background model, is obtained by modeling
Prospect bianry image and background image；
C shadow removals, the shade of prospect bianry image is removed；
D mass detections, obtain position and the size of agglomerate.
3. a kind of multiple target vision-based detection according to claim 2 and tracking, it is characterised in that the shadow removal
Method using the method based on model or the method based on attribute；Based on the method for model be by using scene, target and
Shadow model is set up in illumination, and rib, angle to three-dimensional body, line are matched；It is the geometry using shade based on the method for attribute
Feature, brightness, color information are indicating shadow region.
4. a kind of multiple target vision-based detection according to claim 2 and tracking, it is characterised in that the mass detection
Use the shadow detection method of Rita Cucchiara.
5. a kind of multiple target vision-based detection according to claim 1 and tracking, it is characterised in that in the step 4
Target following model adopts Kalman filter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610946656.1A CN106530325A (en) | 2016-10-26 | 2016-10-26 | Multi-target visual detection and tracking method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610946656.1A CN106530325A (en) | 2016-10-26 | 2016-10-26 | Multi-target visual detection and tracking method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106530325A true CN106530325A (en) | 2017-03-22 |
Family
ID=58292870
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610946656.1A Pending CN106530325A (en) | 2016-10-26 | 2016-10-26 | Multi-target visual detection and tracking method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106530325A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798686A (en) * | 2017-09-04 | 2018-03-13 | 华南理工大学 | A kind of real-time modeling method method that study is differentiated based on multiple features |
| CN112465861A (en) * | 2020-11-19 | 2021-03-09 | 西北工业大学 | Relevant filtering visual target tracking method based on self-adaptive mask |
| CN114693702A (en) * | 2022-03-24 | 2022-07-01 | 小米汽车科技有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
| CN112465861B (en) * | 2020-11-19 | 2024-05-10 | 西北工业大学 | Relevant filtering visual target tracking method based on self-adaptive mask |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739686A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Moving object tracking method and system thereof |
| CN101887587A (en) * | 2010-07-07 | 2010-11-17 | 南京邮电大学 | Multi-target track method based on moving target detection in video monitoring |
-
2016
- 2016-10-26 CN CN201610946656.1A patent/CN106530325A/en active Pending
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101739686A (en) * | 2009-02-11 | 2010-06-16 | 北京智安邦科技有限公司 | Moving object tracking method and system thereof |
| CN101887587A (en) * | 2010-07-07 | 2010-11-17 | 南京邮电大学 | Multi-target track method based on moving target detection in video monitoring |
Non-Patent Citations (1)
| Title |
|---|
| 孙华东: "多目标视觉检测与跟踪方法研究及视频监控软件平台的开发", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107798686A (en) * | 2017-09-04 | 2018-03-13 | 华南理工大学 | A kind of real-time modeling method method that study is differentiated based on multiple features |
| CN112465861A (en) * | 2020-11-19 | 2021-03-09 | 西北工业大学 | Relevant filtering visual target tracking method based on self-adaptive mask |
| CN112465861B (en) * | 2020-11-19 | 2024-05-10 | 西北工业大学 | Relevant filtering visual target tracking method based on self-adaptive mask |
| CN114693702A (en) * | 2022-03-24 | 2022-07-01 | 小米汽车科技有限公司 | Image processing method, image processing device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10956756B2 (en) | Hazard detection from a camera in a scene with moving shadows | |
| Park et al. | Continuous localization of construction workers via integration of detection and tracking | |
| Møgelmose et al. | Trajectory analysis and prediction for improved pedestrian safety: Integrated framework and evaluations | |
| CN105405154B (en) | Target object tracking based on color-structure feature | |
| Veeraraghavan et al. | Computer vision algorithms for intersection monitoring | |
| US9111444B2 (en) | Video and lidar target detection and tracking system and method for segmenting moving targets | |
| CN104091348A (en) | Multi-target tracking method integrating obvious characteristics and block division templates | |
| Li et al. | Multiple lane boundary detection using a combination of low-level image features | |
| CN102243765A (en) | Multi-camera-based multi-objective positioning tracking method and system | |
| Moghadam et al. | Road direction detection based on vanishing-point tracking | |
| CN103729861A (en) | Multiple object tracking method | |
| Tsalatsanis et al. | Vision based target tracking and collision avoidance for mobile robots | |
| Hossain et al. | Fast-D: When non-smoothing color feature meets moving object detection in real-time | |
| Kanhere et al. | Vehicle segmentation and tracking in the presence of occlusions | |
| Qing et al. | A novel particle filter implementation for a multiple-vehicle detection and tracking system using tail light segmentation | |
| CN102810206A (en) | Real-time loitering detection method based on dynamic programming | |
| CN106530325A (en) | Multi-target visual detection and tracking method | |
| Ling et al. | Colour-based object tracking in surveillance application | |
| Luber et al. | Learning to detect and track people in rgbd data | |
| Michael et al. | Fast change detection for camera-based surveillance systems | |
| Kim et al. | Traffic Accident Detection Based on Ego Motion and Object Tracking | |
| Huang et al. | A vehicle flow counting system in rainy environment based on vehicle feature analysis. | |
| Denman et al. | Group segmentation during object tracking using optical flow discontinuities | |
| Mueller et al. | Continuous stereo self-calibration on planar roads | |
| Bui et al. | Multi-sensors people detection system for heavy machines |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20170322 |
|
| RJ01 | Rejection of invention patent application after publication |