CN101996307A - Intelligent video human body identification method - Google Patents
Intelligent video human body identification method Download PDFInfo
- Publication number
- CN101996307A CN101996307A CN2009100561861A CN200910056186A CN101996307A CN 101996307 A CN101996307 A CN 101996307A CN 2009100561861 A CN2009100561861 A CN 2009100561861A CN 200910056186 A CN200910056186 A CN 200910056186A CN 101996307 A CN101996307 A CN 101996307A
- Authority
- CN
- China
- Prior art keywords
- human body
- frame
- human motion
- image
- human
- 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 provides an intelligent video human body identification method belonging to the technical field of computer graphics. The method comprises the following steps of: (1) establishing a static background model; (2) determining a human body movement outline and a background subtraction by utilizing interframe difference to obtain a human body movement areas, and filling the human body movement areas into the human body movement outline; (3) carrying out binaryzation on a filling result obtained in the step (2) by utilizing a maximum variance threshold segmentation method, and meanwhile analyzing the segmentation validity; and (4) for the valid segmentation, merging the human body movement areas, tracking the human body movement locus, updating the static background model and returning to the step (2). The invention has good identification effect and high robustness and does not need manual intervention. Moreover, the invention has strong interference resistance, can accurately eliminate the condition like the sudden change of the indoor light and can also filter off the interference of leaf swing and air flyers and the common outdoor environments like the sudden change of the weather, thereby reducing the false tracking phenomenon and false alarm phenomenon.
Description
Technical field
The invention belongs to the computer graphic image field that learns a skill, is the recognition methods of a kind of intelligent video human body.
Background technology
Intelligent video analysis is that computer vision methods is incorporated in the video monitoring, this technology comprises the content of automatically being carried out aspects such as motion target detection, target following, target classification and behavior understanding by sequence of video images, purpose is to set up mapping relations between image and picture are described, thereby makes computing machine can analyze and understand content in the video pictures.The intelligent video technology of being mentioned in the video monitoring mainly refers to the key message in automatic analysis and the extraction video source, by the powerful data processing function of computing machine, mass data in the video pictures is carried out high speed analysis, filter out the unconcerned information of supervisor, only provide the key message of usefulness for the supervisor.Wherein, the detection of human body and identification are most important parts in the intelligent video technology.
The identification of intelligent video human body need have stronger antijamming capability, effectively solves the DE Camera Shake that causes owing to external environment, noise that light and weather cause or the like.At present, several moving target detecting methods commonly used have frame-to-frame differences method (Temporal Difference), background subtraction method (BackgroundSubtraction) and optical flow method (Optical Flow).The motion detection of optical flow method has adopted the time dependent light stream characteristic of dynamic object, and computing method are quite complicated, abandoned by industry in the environment that real-time video is handled.
The frame-to-frame differences method is time differencing method again, and it has made full use of the feature of video image, extracts needed dynamic object information from the video flowing that obtains continuously.The video image of Cai Jiing in the ordinary course of things, if adjacent two frames of careful contrast, can find that wherein most background pixel all remains unchanged, only the pixel diversity ratio at the part consecutive frame that the present image moving target is arranged is bigger, and the time difference method utilizes subtracting each other of consecutive frame image to extract the information of present image moving target.Though the time difference method for testing motion has stronger adaptivity for dynamic environment, generally can not extract all relevant feature pixels fully, be easy to generate cavitation in movement entity inside, be merely able to detect the edge of target.And when the moving target stop motion, general time difference method just lost efficacy.
The background subtraction method is to utilize the difference of present image and background image to detect moving target.It is a kind of method the most frequently used in the present motion detection, characteristic than more comprehensive moving target generally can be provided comparatively speaking, most researchist is devoted to develop practical more background model at present, is used to reduce the influence of dynamic scene variation for the moving object detection effect.The simplest background model is the time average image, promptly utilizes Same Scene in the average image of the period background model as this scene.The modeling of background is the key problem in technology of background subtraction method.But,,, shake with the wind such as sunniness direction, shadow, leaf etc. in case after setting up, all compare sensitivity for any variation that this scene image took place because this method background model is fixed.
Therefore, the present invention is necessary to provide a kind of new method to solve the problems referred to above.
Summary of the invention
In order to overcome the shortcoming of above-mentioned classic method, the invention provides the recognition methods of a kind of intelligent video human body, can effectively improve the accuracy of identification.
The present invention solves above-mentioned technical matters by such technical scheme:
The recognition methods of a kind of intelligent video human body, this method comprises the steps:
The first step is set up the static background model;
In second step, utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone, and the human motion zone is filled in the human motion profile;
In the 3rd step, utilize the maximum variance threshold segmentation method that the second step gained is filled the result and carry out binaryzation, simultaneously the validity of cutting apart;
The 4th step for effectively cutting apart, merged the human motion zone, and followed the tracks of human body motion track, upgraded the static background model, and returned for second step.
As a kind of improvement of the present invention, pass through the abstraction sequence frame in the first step, and after carrying out medium filtering on the time domain, obtain the static background model at sequence frame.
As a kind of improvement of the present invention, in the 3rd step: note t is the segmentation threshold of human body moving region and background image, is worth g=w when t makes
0* (u
0-u)
2+ w
1* (u
1-u)
2When maximum, t is the maximum variance threshold value of cutting apart; Wherein, to account for image scaled be w to the pixel number in human motion zone
0, average gray is u
0It is w1 that the background image pixel number accounts for image scaled, and average gray is u
1, then the overall average gray scale of image is: u=w
0* u
0+ w
1* u
1
As a kind of improvement of the present invention, if variance g less than an empirical value, i.e. human motion zone and background image can't significantly be distinguished, directly the pending image with human motion zone and background image formation all is set to 0; Wherein the span of empirical value is between 100 to 200.
As a kind of improvement of the present invention, described method comprises the steps:
Step 11, from video source abstraction sequence frame;
Step 12, transfer the sequence frame that extracts to gray level image, and do smothing filtering;
Step 13, present frame carry out luminance standardization;
Step 14, utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone;
Step 15, the human motion zone is filled in the human motion profile;
Step 16, utilize the maximum variance threshold segmentation method that the gained result is carried out binaryzation;
Step 17, judge cut apart whether effective, if judged result is for being that then execution in step 18;
Step 18, find the human motion profile, merge the human motion zone;
Step 19, follow the tracks of and draw human body motion track;
Step 20, renewal static background model;
Step 21, judge whether last frame, if not, step 11 then returned, if finish this identification.
As a kind of improvement of the present invention, whether the sequence frame that extracts in the determining step 12 is preceding 25 frames, if judged result is for being then to carry out following steps:
Step 121, transfer preceding 25 frame pictures to gray level image, deposit buffering in;
Step 122, then preceding 25 frame pictures are carried out medium filtering on time domain,, and return step 11 as the static background model.
As a kind of improvement of the present invention, in the step 17, if judged result is not, then the pending image with human motion zone and background image formation all is set to 0, and returns step 11.
The present invention compares with existing method, its advantage is, recognition effect is good, robustness is high, need not manual intervention, in addition, this method strong interference immunity, can eliminate situations such as indoor light sudden change accurately, also can filter swing, the airflight thing of leaf interference, rain, snow, haze and common outdoor environment such as change in weather, thereby reduced common tracking error phenomenon and false alarm situation in the industry.
Description of drawings
In order to be illustrated more clearly in the embodiment of the invention or technical scheme of the prior art, to do to introduce simply to the accompanying drawing of required use in embodiment or the description of the Prior Art below, apparently, accompanying drawing in describing below only is some embodiments of the present invention, for those of ordinary skills, under the prerequisite of not paying creative work, can also obtain other accompanying drawing according to these accompanying drawings.
Fig. 1 is the overview flow chart of intelligent video human body of the present invention recognition methods.
Fig. 2 is the detail flowchart of intelligent video human body of the present invention recognition methods.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the invention, the technical scheme in the embodiment of the invention is clearly and completely described.
The invention provides the recognition methods of a kind of intelligent video human body.See also Fig. 1, described method comprises the steps:
The first step is set up the static background model, at first by the sequence frame about 1 second general 25 frames is obtained the static background model after carrying out medium filtering on the time domain, thereby has eliminated the single frames instability of model as a setting;
In second step, utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone, and the human motion zone is filled in the human motion profile;
In the 3rd step, utilize the maximum variance Threshold Segmentation that the second step gained is filled the result and carry out binaryzation, simultaneously the validity of cutting apart;
The 4th step for effectively cutting apart, merged the human motion zone, and realized the human body motion track tracking, upgraded the static background model, and returned for second step.
Below respectively the enforcement of each step is described in detail:
1. static background modelling and renewal
The foundation of static background model is the key problem in technology of background subtraction method, can carry out the time domain medium filtering by the sequence frame to about 1 second general 25 frames and obtain.Setting up good static background model all compares responsive for any variation that this scene image took place, therefore the present invention utilizes and follows the tracks of the back result, utilize formula B (x) to upgrade the static background model, it can be reduced because the scene that noise causes changes the influence for the moving object detection effect.
2. utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone, and the human motion zone is filled in the human motion profile;
Utilize the frame-to-frame differences method to obtain the human motion profile, extract the whole human body moving region with the background subtraction method then.
For the video of camera acquisition, modern I
n(x) be location of pixels x, the brightness value during time t=n.The three-frame difference method at first is that the acquiescence pixel is legal motion, if the brightness value of a pixel between present frame and previous frame and present frame and on interframe alter a great deal, judge that then this pixel moves, that is:
When | I
n(x)-I
N-1(x) |＞T
n(x) and | I
n(x)-I
N-2(x) |＞T
n(x) time, the judgement pixel is moved.
Here, T
n(x) be an effective brightness change threshold of describing pixel x place, this threshold value obtains by the time domain statistic law.The subject matter that the frame-to-frame differences method exists is that the same pixel of brightness of target internal is not included in the set of pixels of " motion ".
The background subtraction method obtains human motion zone b
n
With b
nIn the human motion profile that cluster obtains to the frame-to-frame differences method, promptly can obtain a comparatively complete human motion zone.
B (x) and T (x) are along with the time is brought in constant renewal in:
Here α is a time constant, is used to specify the speed that background information is upgraded, and is an empirical value; M is a filter factor.
Here being noted that each value does not have only when pixel is judged as when moving just can change, just the part of static background.If each location of pixels that does not move is from time series, B
n(x) be similar with the local zone time average, T
n(x) be doubly similar with the m of local zone time normal brightness difference.B
n(x) and T
n(x) all be to calculate with unlimited pulse feedback (IIR) wave filter to obtain.
3. maximum variance Threshold Segmentation and cut apart availability deciding
3.1, the maximum variance Threshold Segmentation, be the simple high efficiency method that self-adaptation is calculated single threshold.
To a secondary gray level image, note t is the segmentation threshold of human body moving region and background image, and it is w that the pixel number in human motion zone accounts for image scaled
0, average gray is u
0, it is w1 that the background image pixel number accounts for image scaled, average gray is u
1, then the overall average gray scale of image is: u=w
0* u
0+ w
1* u
1
From the minimum gradation value to the maximum gradation value, travel through t, be worth g=w when t makes
0* (u
0-u)
2+ w
1* (u
1-u)
2T is the optimal threshold of cutting apart when maximum, that is to say the maximum variance threshold values.
Human motion zone and background image two parts that maximum variance threshold value t is partitioned into have constituted entire image, and human motion zone value u
0, probability is w
0, background image value u
1, probability is w
1, grand mean is u, variance yields is g.
3.2, judge whether current Threshold Segmentation effective
Part human body moving region mistake is divided into background image or part background image mistake and is divided into the human motion zone and can causes all that difference diminishes between human motion zone and the background image.And variance is as the inhomogeneity tolerance of intensity profile, and variance yields is big more, illustrates that to constitute between human motion zone and the background image difference big more, if variance yields maximum, then mean the misclassification probability minimum, therefore, whether variance yields can be used to characterize current Threshold Segmentation effective.
If variance is fuzzy less than an empirical value explanation human motion zone and background image, differentiation that can't be clearly, just mean that optimal threshold is nonsensical in fact, can judge in the scene it is undesired signal so, at this moment can directly the human motion zone all be set to 0.If variance, illustrates then that current Threshold Segmentation is effective greater than empirical value, then carried out for the 4th step, the human motion zone is merged.Wherein the scope of empirical value is between 100 to 200.
This method has solved the interference that causes owing to the interference of the swing of indoor light sudden change, leaf, airflight thing, rain, snow, haze etc. effectively.
4. the human motion zone merges and the human body motion track tracking
4.1, can obtain effective segmentation result through above-mentioned steps, but not that each human body only can obtain an effective exercise zone in the ideal owing to cut apart, a human body is split into a plurality of effective exercises zone often, but these effective exercise zones are adjacent near especially, and, carry out the merging and the cluster of target area according to the tracking results of former frame.
Combining step is as follows:
Step 1 is carried out cluster according to certain threshold value with the bounding box in approaching target zone and is merged, and wherein bounding box is to surround each divided effective exercise zone;
Step 2 is if certain bounding box in the present frame in the bounding box inside that previous frame merged, directly merges;
Step 3 is if not in the bounding box inside that merged, then carry out cluster according to threshold value; If current block is nearer from current new merging bounding box, merge so; Do not close on yet merging if merge piece with previous frame; Otherwise independently be that a bounding box is labeled as merging;
Step 4 is upgraded the result that bounding box merges, and marks which bounding box and merged;
Step 5 for all bounding box repeating steps 3,4,5, all merges up to all bounding boxs.
In previous frame,, illustrate that then the previous frame ratio of division is more satisfactory if do not merge piece; If it is undesirable that previous frame is cut apart, then to carry out merging, rule of thumb threshold value is judged.
4.2, realize and draw human body motion track by kalman filter method and follow the tracks of, kalman filter method is to be the optimum criterion of estimating with the least mean-square error, adopt the state-space model of signal and noise, utilize the observed reading of the estimated value of previous moment and current time to upgrade estimation, obtain the estimated value of current time state variable.Kalman filtering is suitable for handling in real time and Computing, can effectively follow the tracks of human body motion track.
See also Fig. 2, intelligent video human body provided by the invention recognition methods specifically comprises the steps:
Step 11, from video source, camera or video file abstraction sequence frame;
Step 12, judge whether it is preceding 25 frames, if, execution in step 121 and step 122, if not, execution in step 13;
Step 121, transfer preceding 25 frame pictures to gray level image, deposit buffering in;
Step 122, then preceding 25 frame pictures are carried out medium filtering on time domain,, and return step 11 as the static background model;
Step 13, transfer preceding 25 frame pictures to gray level image, and do smothing filtering;
Step 14, present frame carry out luminance standardization;
Step 15, utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone;
Step 16, the human motion zone is filled in the human motion profile;
Step 17, maximum variance Threshold Segmentation;
Step 18, judge whether segmentation result effective, if judged result for not, then execution in step 181; If judged result is for being that then execution in step 19;
Step 181, the pending image setting that human motion zone and background image are constituted are 0, and return step 11;
Step 19, find the human motion profile, merge the human motion zone;
Step 20, follow the tracks of and draw human body motion track;
Step 21, renewal static background model;
Step 22, judge whether last frame, if not, step 11 then returned, if finish this identification.
The present invention can solve the problem that industry is mentioned effectively, weather conditions such as the rain of for example indoor light sudden change, outdoor environment, snow, mist, also can filter the swing of leaf, the interference of airflight thing, thereby improve tracking performance, reduce situations such as tracking error false alarm.
The fusion application of frame-to-frame differences method and background subtraction method has been solved the incomplete shortcoming of traditional algorithm movement human information extraction, with the possibility of maximum movement human information completely profile is offered the user so that the human body motion track of back is followed the tracks of and the static background renewal.
The threshold value availability deciding can judge very accurately whether current scene has real detection target to occur, this method can improve the tolerance of intelligent video monitoring system for weather such as sleet on the one hand very effectively, also can eliminate some airflight things on the other hand, the passive shake of camera etc. are factor inevitably.
In real time the static background update algorithm has solved the fixed background model and all compare the shortcoming of sensitivity for any variation that scene image took place, and such as some unavoidable spontaneous phenomenons: solar radiation, shadow, leaf are shaken etc. with the wind.
The above; only be the specific embodiment of the present invention, but protection scope of the present invention is not limited thereto, anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; can expect easily changing or replacing, all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion by described protection domain with claim.
Claims (7)
1. intelligent video human body recognition methods is characterized in that this method comprises the steps:
The first step is set up the static background model;
In second step, utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone, and the human motion zone is filled in the human motion profile;
In the 3rd step, utilize the maximum variance threshold segmentation method that the second step gained is filled the result and carry out binaryzation, simultaneously the validity of cutting apart;
The 4th step for effectively cutting apart, merged the human motion zone, and followed the tracks of human body motion track, upgraded the static background model, and returned for second step.
2. intelligent video human body according to claim 1 recognition methods is characterized in that, passes through the abstraction sequence frame in the first step, and obtains the static background model at sequence frame after carrying out medium filtering on the time domain.
3. intelligent video human body according to claim 1 recognition methods is characterized in that, in the 3rd step: note t is the segmentation threshold of human body moving region and background image, is worth g=w when t makes
0* (u
0-u)
2+ w
1* (u
1-u)
2When maximum, t is the maximum variance threshold value of cutting apart; Wherein, to account for image scaled be w to the pixel number in human motion zone
0, average gray is u
0It is w1 that the background image pixel number accounts for image scaled, and average gray is u
1, then the overall average gray scale of image is: u=w
0* u
0+ w
1* u
1
4. intelligent video human body according to claim 3 recognition methods, it is characterized in that, if variance g is less than an empirical value, i.e. human motion zone and background image can't significantly be distinguished, and directly the pending image with human motion zone and background image formation all is set to 0; Wherein the span of empirical value is between 100 to 200.
5. intelligent video human body according to claim 1 recognition methods is characterized in that described method further comprises the steps:
Step 11, from video source abstraction sequence frame;
Step 12, transfer the sequence frame that extracts to gray level image, and do smothing filtering;
Step 13, present frame carry out luminance standardization;
Step 14, utilize the frame-to-frame differences method to determine that human motion profile and background subtraction method obtain the human motion zone;
Step 15, the human motion zone is filled in the human motion profile;
Step 16, utilize the maximum variance threshold segmentation method that the gained result is carried out binaryzation;
Step 17, judge cut apart whether effective, if judged result is for being that then execution in step 18;
Step 18, find the human motion profile, merge the human motion zone;
Step 19, follow the tracks of and draw human body motion track;
Step 20, renewal static background model;
Step 21, judge whether it is last frame, if not, step 11 then returned, if finish this identification.
6. intelligent video human body according to claim 5 recognition methods is characterized in that, whether the sequence frame that extracts in the determining step 12 is preceding 25 frames, if judged result is for being then to carry out following steps:
Step 121, transfer preceding 25 frame pictures to gray level image, deposit buffering in;
Step 122, then preceding 25 frame pictures are carried out medium filtering on time domain,, and return step 11 as the static background model.
7. intelligent video human body according to claim 5 recognition methods is characterized in that, in the step 17, if judged result is not, then the pending image with human motion zone and background image formation all is set to 0, and returns step 11.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100561861A CN101996307A (en) | 2009-08-10 | 2009-08-10 | Intelligent video human body identification method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009100561861A CN101996307A (en) | 2009-08-10 | 2009-08-10 | Intelligent video human body identification method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101996307A true CN101996307A (en) | 2011-03-30 |
Family
ID=43786446
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009100561861A Pending CN101996307A (en) | 2009-08-10 | 2009-08-10 | Intelligent video human body identification method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101996307A (en) |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102521582A (en) * | 2011-12-28 | 2012-06-27 | 浙江大学 | Human upper body detection and splitting method applied to low-contrast video |
| CN102625031A (en) * | 2012-03-21 | 2012-08-01 | 天津天地伟业数码科技有限公司 | Mid-value filter de-noising system and mid-value filter de-noising method for video data |
| CN103002195A (en) * | 2011-09-09 | 2013-03-27 | 中怡(苏州)科技有限公司 | Motion detecting method and motion detecting device |
| CN103208005A (en) * | 2012-01-13 | 2013-07-17 | 富士通株式会社 | Object recognition method and object recognition device |
| CN104063883A (en) * | 2014-07-07 | 2014-09-24 | 杭州银江智慧医疗集团有限公司 | Surveillance video abstract generating method based on combination of object and key frames |
| CN105654647A (en) * | 2016-01-28 | 2016-06-08 | 中北大学 | Identification method for judging home invasion in real time |
| CN107742306A (en) * | 2017-09-20 | 2018-02-27 | 徐州工程学院 | Moving Target Tracking Algorithm in a kind of intelligent vision |
| CN108921867A (en) * | 2018-06-01 | 2018-11-30 | 天津煋鸟科技有限公司 | A kind of motion detection and human body recognition method of panoramic picture |
| CN112004056A (en) * | 2020-08-06 | 2020-11-27 | 武汉倍特威视系统有限公司 | Intelligent video analysis method with strong anti-interference capability |
| CN113344967A (en) * | 2021-06-07 | 2021-09-03 | 哈尔滨理工大学 | Dynamic target identification tracking method under complex background |
-
2009
- 2009-08-10 CN CN2009100561861A patent/CN101996307A/en active Pending
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103002195A (en) * | 2011-09-09 | 2013-03-27 | 中怡(苏州)科技有限公司 | Motion detecting method and motion detecting device |
| CN103002195B (en) * | 2011-09-09 | 2016-03-30 | 中磊电子（苏州）有限公司 | Method for testing motion and motion detection apparatus |
| CN102521582A (en) * | 2011-12-28 | 2012-06-27 | 浙江大学 | Human upper body detection and splitting method applied to low-contrast video |
| CN103208005A (en) * | 2012-01-13 | 2013-07-17 | 富士通株式会社 | Object recognition method and object recognition device |
| CN102625031A (en) * | 2012-03-21 | 2012-08-01 | 天津天地伟业数码科技有限公司 | Mid-value filter de-noising system and mid-value filter de-noising method for video data |
| CN104063883B (en) * | 2014-07-07 | 2018-03-16 | 杭州银江智慧医疗集团有限公司 | A kind of monitor video abstraction generating method being combined based on object and key frame |
| CN104063883A (en) * | 2014-07-07 | 2014-09-24 | 杭州银江智慧医疗集团有限公司 | Surveillance video abstract generating method based on combination of object and key frames |
| CN105654647A (en) * | 2016-01-28 | 2016-06-08 | 中北大学 | Identification method for judging home invasion in real time |
| CN107742306A (en) * | 2017-09-20 | 2018-02-27 | 徐州工程学院 | Moving Target Tracking Algorithm in a kind of intelligent vision |
| CN107742306B (en) * | 2017-09-20 | 2021-03-23 | 徐州工程学院 | Moving target tracking algorithm in intelligent vision |
| CN108921867A (en) * | 2018-06-01 | 2018-11-30 | 天津煋鸟科技有限公司 | A kind of motion detection and human body recognition method of panoramic picture |
| CN112004056A (en) * | 2020-08-06 | 2020-11-27 | 武汉倍特威视系统有限公司 | Intelligent video analysis method with strong anti-interference capability |
| CN113344967A (en) * | 2021-06-07 | 2021-09-03 | 哈尔滨理工大学 | Dynamic target identification tracking method under complex background |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101996307A (en) | Intelligent video human body identification method | |
| CN102222214A (en) | Fast object recognition algorithm | |
| CN101739686B (en) | Moving object tracking method and system thereof | |
| US20230289979A1 (en) | A method for video moving object detection based on relative statistical characteristics of image pixels | |
| CN101739551B (en) | Method and system for identifying moving objects | |
| CN100545867C (en) | Aerial shooting traffic video frequency vehicle rapid checking method | |
| CN105427626B (en) | A kind of statistical method of traffic flow based on video analysis | |
| CN112036254B (en) | Moving vehicle foreground detection method based on video image | |
| CN111401284B (en) | Door opening and closing state identification method based on image processing | |
| CN106022231A (en) | Multi-feature-fusion-based technical method for rapid detection of pedestrian | |
| CN104616290A (en) | Target detection algorithm in combination of statistical matrix model and adaptive threshold | |
| CN101621615A (en) | Self-adaptive background modeling and moving target detecting method | |
| CN109684996A (en) | Real-time vehicle based on video passes in and out recognition methods | |
| CN106778540B (en) | Parking detection is accurately based on the parking event detecting method of background double layer | |
| CN108022249A (en) | A kind of remote sensing video satellite moving vehicle target region of interest extraction method | |
| CN103945089A (en) | Dynamic target detection method based on brightness flicker correction and IP camera | |
| Lian et al. | A novel method on moving-objects detection based on background subtraction and three frames differencing | |
| CN108920997A (en) | Judge that non-rigid targets whether there is the tracking blocked based on profile | |
| CN102902951A (en) | System and method for vehicle target location and event detection on basis of high-definition video monitoring images | |
| CN110728700B (en) | Moving target tracking method and device, computer equipment and storage medium | |
| Tao et al. | Real-time detection and tracking of moving object | |
| CN107301655B (en) | Video moving target detection method based on background modeling | |
| CN107886524B (en) | Method for identifying motion track of elevator door | |
| Sutjiadi et al. | Adaptive background extraction for video based traffic counter application using Gaussian mixture models algorithm | |
| abd el Azeem Marzouk | Modified background subtraction algorithm for motion detection in surveillance systems |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110330 |