CN106485729A - A kind of moving target detecting method based on mixed Gauss model - Google Patents
A kind of moving target detecting method based on mixed Gauss model Download PDFInfo
- Publication number
- CN106485729A CN106485729A CN201610858219.4A CN201610858219A CN106485729A CN 106485729 A CN106485729 A CN 106485729A CN 201610858219 A CN201610858219 A CN 201610858219A CN 106485729 A CN106485729 A CN 106485729A
- Authority
- CN
- China
- Prior art keywords
- region
- variation
- image
- moving target
- pixel
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
Landscapes
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
Abstract
The invention discloses a kind of moving target detecting method based on mixed Gaussian, comprises the following steps：1）Set up mixed Gauss model；2）Region of variation is obtained by difference, and judgement is measured and is analyzed to region of variation area；3）Analysis judges whether region of variation there occurs light sudden change；4）According to step 3）Judged result adopt different moving target detecting methods；5）Detect moving target.Whether the Background Reconstruction mechanism that the present invention is updated using cycle period, by measurement and judgement to region of variation area, and there occurs the judgement of light sudden change, optionally using background subtraction or frame differential method detection moving target to region of variation.The present invention preferably overcomes traditional background subtraction and compares sensitive issue to illumination variation, while also solve the problems, such as easily to produce the diplopia of long period when changing between prospect and background in scene, with higher robustness.
Description
Technical field
The invention belongs to technical field of video monitoring, is related to a kind of moving object detection side based on mixed Gauss model
Method.
Background technology
Intelligent video monitoring is referred in the case of human intervention is not needed, using computer vision analysis method to video
Sequence is automatically analyzed, and realizes moving object detection, classification, identification, tracking etc., and on this basis, by presetting
Rule the behavior of target is analyzed, so as to for the offer reference that takes further measures.Wherein, the purpose of motion detection is
By the analysis to monitor video image sequence, determine in monitoring scene and moving target is had or not, and then moving region from detection
Extract in image.It is to carry out motion target tracking, classification and identification etc. subsequently to locate to moving region accurate and effective Ground Split
The basic premise of reason.
At present, moving target detecting method mainly has optical flow method, frame differential method, background subtraction.And it is solid in video camera
In fixed outdoor video monitoring system, most common method is the background subtraction based on mixed Gauss model, and the method is every
Individual pixel sets up multiple Gauss models, and the adaptivity to background is high, can preferably describe complex background.But mixed Gaussian
When model becomes foreground moving object for long-time stagnation object by background, background model renewal speed is slower；When object by
Static when switching to motion and fleeing from background, may miss and background be appeared area prospect is detected as, occur foreground object in testing result
" shadow ".In addition, based on the background subtraction on the basis of mixed Gauss model, be based on fixation, static background, but real
Border environment is always complicated and changeable, for example slight jitter of illumination variation, background perturbation and camera etc., can all affect motion
The sensitivity of target detection and accuracy.
Content of the invention
For mixed Gauss model and the above-mentioned deficiency of background subtraction, it is proposed that a kind of by mixed Gauss model and interframe
The moving object detection algorithm that calculus of finite differences combines.
The moving target detecting method based on mixed Gauss model that the present invention is adopted, technical scheme are as follows：
Step 1：Using mixed Gauss model, initial background, and the Background Reconstruction mechanism updated using cycle period, with continuous
Update background；
Step 2：Current image frame is read in the video shot from video camera；
Step 3：Region of variation is obtained with previous background image difference to current frame image；
Step 4：Region of variation area described in step 3 is measured；
Step 5：Analysis judges whether described region of variation area exceedes threshold value, only when the area of region of variation is entirely being schemed
Accounting in picture just thinks there is moving target when exceeding given threshold, then carries out the process of next step to it；
Step 6：Analysis judges whether region of variation there occurs that light changes；
Step 7：If region of variation does not occur light to change, just moving target is detected using background subtraction；If change
There is light change in region, just detect moving target using frame differential method；
Step 8：Detect moving target.
The beneficial effects of the present invention is:Propose one kind mixed Gauss model background method is combined with frame differential method
Moving object detection innovatory algorithm, preferably overcome traditional background subtraction and sensitive issue compared to illumination variation, with
When also solve the problems, such as when changing between prospect and background in scene easily to produce the diplopia of long period.The method can
So that the detection of moving target under light situation of change is tackled, with higher robustness, the change of real background can be responded rapidly to
Change, the background of generation can accurately reflect scene information, be the motion target tracking of next step and identification provide reliable according to
According to.
Description of the drawings
In order to be illustrated more clearly that the embodiment of the present invention or technical scheme of the prior art, below will be to embodiment or existing
Accompanying drawing to be used needed for technology description is had to be briefly described, it should be apparent that, drawings in the following description are only this
Some embodiments of invention, for those of ordinary skill in the art, on the premise of not paying creative work, acceptable
Other accompanying drawings are obtained according to these accompanying drawings.
Fig. 1 is moving object detection algorithm flow chart of the present invention.
Fig. 2 is the measuring method flow chart of region of variation area of the present invention.
Fig. 3 is region of variation light sudden change overhaul flow chart of the present invention.
Specific embodiment
With reference to Figure of description, a kind of moving target detecting method provided in an embodiment of the present invention is embodied as
Mode is illustrated.
【Step S501】：Set up mixed Gauss model；
This step sets up mixed Gauss model, the Background Reconstruction mechanism updated using cycle period, and specific algorithm is：
Hypothesis cycle period is T, and pixel average is, pixel undulating value be, pixel prospect number be, grey scale pixel value be；Represent different threshold values；
1）Certain pixel is assumed for R (x, y), judge whether this needs the algorithm being updated to be in t：
Just illustrate that when R (x, y)=1 mixed Gauss model at this needs to rebuild, whereinThe value of setting is not less than O.7；If
O.4 the value that puts is not more than；
2）As t=T, a cycle period terminates, then pixel average, pixel undulating value, pixel prospect
NumberMore new algorithm be：
;
;
.
【Step S502】：Current image frame is read in the video shot from video camera.
【Step S503】：Region of variation is obtained with previous background image difference to current frame image.
【Step S504】：Described region of variation area is measured.
Measuring method described in this step to region of variation area is：
【Step S1】：Obtain image change region；
【Step S2】：Image to image change region carries out binary conversion treatment；
【Step S3】：Image to image change region is progressively scanned；
【Step S4】：Count the number of pixels that pixel value in the region of variation image is 1；
【Step S5】：With the number of pixels described in step S4 divided by the pixel sum of entire image, region of variation can be obtained
Account for the approximate ratio of total image area.
【Step S505】：Analysis judges whether described region of variation area exceedes threshold value
Approximate ratio described in step S504 compared with the threshold value for setting, if it exceeds during given threshold, then it is assumed that there is motion mesh
Mark, carries out subsequent treatment.If be not above given threshold, then it is assumed that without moving target, return to the new image for obtaining
Region of variation is detected.
【Step S506】：Analysis judges whether described region of variation there occurs light sudden change；
Binary conversion treatment is carried out to the region of variation image, counts the pixel that pixel value in the region of variation image is 1
Number, then total divided by the pixel of entire image with the number of pixels, when the number of the pixel that image pixel value is 1 accounts for pixel
When the percentage of sum exceedes given threshold, then it is assumed that there occurs illuminance abrupt variation.
The determination methods of light sudden change whether are occurred to be described in this step to region of variation light：
【Step S10】：Obtain image change region；
【Step S11】：Image to image change region carries out binary conversion treatment；
【Step S12】：Count the number of pixels that pixel value in the region of variation image is 1；
【Step S13】：With the number of pixels described in step S3 divided by the pixel sum of entire image, region of variation can be obtained
Account for the approximate ratio of total image area；
【Step S14】：Approximate ratio described in step S4 compared with the threshold value for setting, if it exceeds during given threshold, then it is assumed that
There occurs illuminance abrupt variation；If be not above given threshold, then it is assumed that illuminance abrupt variation does not occur.
【Step S507】：If described region of variation does not occur light sudden change, just using background subtraction detection fortune
Moving-target；If described region of variation occurs light sudden change, just moving target is detected using frame differential method.
【Step S508】：Detect moving target.
In sum, a kind of moving target detecting method based on mixed Gauss model proposed by the present invention, can tackle
The detection of moving target under light situation of change, can respond rapidly to the change of real background, and the background of generation can accurately be anti-
Scene information is reflected, is that motion target tracking and the identification of next step provides reliable foundation, with higher robustness.
Through the above description of the embodiments, it will be apparent to those skilled in the art, without departing from the present invention
Spirit or essential characteristics in the case of, the present invention can in other forms, structure, arrangement, ratio, and with other components,
Material and part are realizing.In the case of without departing from scope and spirit of the present invention, embodiments disclosed herein can be entered
Other deformation of row and change, if these modifications of the present invention and modification belong to the model of the claims in the present invention and its equivalent technologies
Within enclosing, then the present invention is also intended to comprising these changes and modification.
Claims (4)
1. a kind of moving target detecting method based on mixed Gaussian, it is characterised in that comprise the following steps：
1）Set up mixed Gauss model；
2）Current image frame is read in the video shot from video camera；
3）Region of variation is obtained with previous background image difference to current frame image；
4）Described region of variation area is measured；
5）Analysis judges whether described region of variation area exceedes threshold value；
6）Analysis judges whether described region of variation there occurs light sudden change；
7）If described region of variation does not occur light sudden change, just moving target is detected using background subtraction；If institute
There is light sudden change in the region of variation that states, just detect moving target using frame differential method；
8）Detect moving target.
2. a kind of moving target detecting method based on mixed Gaussian according to claim 1, it is characterised in that described build
Vertical mixed Gauss model step, using following algorithm：
Hypothesis cycle period is T, and pixel average is, pixel undulating value be, pixel prospect number be, grey scale pixel value be；Represent different threshold values；
1）Certain pixel is assumed for R (x, y), judge whether this needs the algorithm being updated to be in t：
Just illustrate that when R (x, y)=1 mixed Gauss model at this needs to rebuild, whereinThe value of setting is not less than O.7；If
O.4 the value that puts is not more than；
2）As t=T, a cycle period terminates, then pixel average, pixel undulating value, pixel prospect
NumberMore new algorithm be：
;
;
.
3. a kind of moving target detecting method based on mixed Gaussian according to claim 1, it is characterised in that the change
Change area measurement to comprise the following steps：
1）Obtain image change region；
2）Image to image change region carries out binary conversion treatment；
3）Image to image change region is progressively scanned；
4）Count the number of pixels that pixel value in the region of variation image is 1；
5）With the number of pixels divided by the pixel sum of entire image, region of variation can be obtained and account for the near of total image area
Like ratio；
6）The approximate ratio compared with the threshold value for setting, if it exceeds during given threshold, then it is assumed that there is moving target, carry out
Subsequent treatment；If be not above given threshold, then it is assumed that without moving target, return to the new image change region for obtaining
Detected.
4. a kind of moving target detecting method based on mixed Gaussian according to claim 1, it is characterised in that analysis is sentenced
The step of whether disconnected described region of variation there occurs light sudden change be：
1）Obtain image change region；
2）Image to image change region carries out binary conversion treatment；
3）Count the number of pixels that pixel value in the region of variation image is 1；
4）With described number of pixels divided by the pixel sum of entire image, region of variation can be obtained and account for total image area
Approximate ratio；
5）The approximate ratio compared with the threshold value for setting, if it exceeds during given threshold, then it is assumed that there occurs illuminance abrupt variation；
If be not above given threshold, then it is assumed that illuminance abrupt variation does not occur.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610858219.4A CN106485729A (en) | 2016-09-29 | 2016-09-29 | A kind of moving target detecting method based on mixed Gauss model |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610858219.4A CN106485729A (en) | 2016-09-29 | 2016-09-29 | A kind of moving target detecting method based on mixed Gauss model |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106485729A true CN106485729A (en) | 2017-03-08 |
Family
ID=58267752
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610858219.4A Pending CN106485729A (en) | 2016-09-29 | 2016-09-29 | A kind of moving target detecting method based on mixed Gauss model |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106485729A (en) |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107333026A (en) * | 2017-06-26 | 2017-11-07 | 电子科技大学 | A kind of monitor video image transmitting being directed under fixed scene and storage method |
| CN108550163A (en) * | 2018-04-19 | 2018-09-18 | 湖南理工学院 | Moving target detecting method in a kind of complex background scene |
| CN108629254A (en) * | 2017-03-24 | 2018-10-09 | 杭州海康威视数字技术股份有限公司 | A kind of detection method and device of moving target |
| CN109798888A (en) * | 2019-03-15 | 2019-05-24 | 京东方科技集团股份有限公司 | Posture determining device, method and the visual odometry of mobile device |
| CN110710194A (en) * | 2019-08-30 | 2020-01-17 | 中新智擎科技有限公司 | Exposure method and device, camera module and electronic equipment |
| CN111444854A (en) * | 2020-03-27 | 2020-07-24 | 科大讯飞(苏州)科技有限公司 | Abnormal event detection method, related device and readable storage medium |
| CN113011219A (en) * | 2019-12-19 | 2021-06-22 | 合肥君正科技有限公司 | Method for automatically updating background in response to light change in occlusion detection |
-
2016
- 2016-09-29 CN CN201610858219.4A patent/CN106485729A/en active Pending
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108629254A (en) * | 2017-03-24 | 2018-10-09 | 杭州海康威视数字技术股份有限公司 | A kind of detection method and device of moving target |
| CN107333026A (en) * | 2017-06-26 | 2017-11-07 | 电子科技大学 | A kind of monitor video image transmitting being directed under fixed scene and storage method |
| CN108550163A (en) * | 2018-04-19 | 2018-09-18 | 湖南理工学院 | Moving target detecting method in a kind of complex background scene |
| CN109798888A (en) * | 2019-03-15 | 2019-05-24 | 京东方科技集团股份有限公司 | Posture determining device, method and the visual odometry of mobile device |
| CN110710194A (en) * | 2019-08-30 | 2020-01-17 | 中新智擎科技有限公司 | Exposure method and device, camera module and electronic equipment |
| WO2021035729A1 (en) * | 2019-08-30 | 2021-03-04 | 中新智擎科技有限公司 | Exposure method and apparatus, image capture module, and electronic device |
| CN110710194B (en) * | 2019-08-30 | 2021-10-22 | 中新智擎科技有限公司 | Exposure method and device, camera module and electronic equipment |
| CN113011219A (en) * | 2019-12-19 | 2021-06-22 | 合肥君正科技有限公司 | Method for automatically updating background in response to light change in occlusion detection |
| CN111444854A (en) * | 2020-03-27 | 2020-07-24 | 科大讯飞(苏州)科技有限公司 | Abnormal event detection method, related device and readable storage medium |
| CN111444854B (en) * | 2020-03-27 | 2021-03-02 | 科大讯飞(苏州)科技有限公司 | Abnormal event detection method, related device and readable storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106485729A (en) | A kind of moving target detecting method based on mixed Gauss model | |
| CN106597463B (en) | Photo-electric proximity sensor and detection method based on dynamic visual sensor chip | |
| CN103295016B (en) | Behavior recognition method based on depth and RGB information and multi-scale and multidirectional rank and level characteristics | |
| CN107133969B (en) | A kind of mobile platform moving target detecting method based on background back projection | |
| CN104978567B (en) | Vehicle checking method based on scene classification | |
| Xiong et al. | Color sensors and their applications based on real-time color image segmentation for cyber physical systems | |
| CN107948465A (en) | A kind of method and apparatus for detecting camera and being disturbed | |
| CN111260693B (en) | High-altitude parabolic detection method | |
| CN102542552B (en) | Frontlighting and backlighting judgment method of video images and detection method of shooting time | |
| CN103325115B (en) | A kind of method of monitoring people counting based on overhead camera head | |
| CN109711332B (en) | Regression algorithm-based face tracking method and application | |
| CN104766342A (en) | Moving target tracking system and speed measuring method based on temporal vision sensor | |
| CN103810722A (en) | Moving target detection method combining improved LBP (Local Binary Pattern) texture and chrominance information | |
| CN110633678B (en) | Quick and efficient vehicle flow calculation method based on video image | |
| CN116309781B (en) | Cross-modal fusion-based underwater visual target ranging method and device | |
| CN109461173A (en) | A kind of Fast Corner Detection method for the processing of time-domain visual sensor signal | |
| CN105469427B (en) | One kind is for method for tracking target in video | |
| CN103489012B (en) | Crowd density detecting method and system based on support vector machine | |
| JP5060264B2 (en) | Human detection device | |
| CN108229524A (en) | A kind of chimney and condensing tower detection method based on remote sensing images | |
| CN112288771B (en) | Method for extracting motion tracks of multiple pig bodies and analyzing behaviors in group environment | |
| CN109635720A (en) | The illegal road occupying real-time detection method actively monitored based on video | |
| CN105469054B (en) | The model building method of normal behaviour and the detection method of abnormal behaviour | |
| CN114972177A (en) | Road disease identification management method and device and intelligent terminal | |
| CN105957060B (en) | A kind of TVS event cluster-dividing method based on optical flow analysis |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20170308 |