CN105046285A - Abnormal behavior identification method based on motion constraints - Google Patents
Abnormal behavior identification method based on motion constraints Download PDFInfo
- Publication number
- CN105046285A CN105046285A CN201510551661.8A CN201510551661A CN105046285A CN 105046285 A CN105046285 A CN 105046285A CN 201510551661 A CN201510551661 A CN 201510551661A CN 105046285 A CN105046285 A CN 105046285A
- Authority
- CN
- China
- Prior art keywords
- optical
- flow feature
- kinematic constraint
- svm
- abnormal behaviour
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Image Analysis (AREA)
Abstract
The invention relates to an abnormal behavior identification method based on motion constraints, and belongs to the field of video monitoring. The abnormal behavior identification method comprises a training process. The training process comprises the following steps that 1, the inputted first set of video frame images are calibrated by adopting a rapid calibration method; 2, background modeling is performed on the calibrated inputted video frame images, and optical flow features are extracted from the frame images by using an optical flow method simultaneously; 3, the optical flow features in the foreground are analyzed and selected by using the motion constraints of pedestrians; and 4, the selected optical flow features are studied by using an SVM so that SVM classifier model parameters which can be used for abnormal behavior identification are obtained. The abnormal behavior identification method also comprises a detection process. In the detection process, a second set of videos are inputted, and the used SVM parameters are obtained in the training process and used for detection and judgment of the model parameters. Optical flow feature calculation is optimized, noise optical flow features are eliminated and the optical flow features of the pedestrians are enabled to be more obvious. Besides, running and other abnormal behaviors can be judged.
Description
Technical field
The present invention relates to a kind of abnormal behaviour discrimination method based on kinematic constraint.
Background technology
When monitoring environment a bit crowded but not serious time, can improve the abnormal behaviour discrimination method based on support vector machine (SVM) based on the work of forefathers, as the tracking of middle employing robust, about following the tracks of 20 people under environment when can block not serious mutually between pedestrian, the method for following the tracks of is carried out in middle proposition for the human body image of defect.
But in environment crowded especially, when mutual coverage extent between pedestrian exceedes certain limit, cannot obtain complete or clearly pedestrian contour feature time, above method will encounter difficulties when understanding crowd behaviour, now needs new method to solve this problem.For this reason, propose the concept of kinematic constraint, and it is combined with the study of Optical-flow Feature, achieve the abnormal behaviour identification under crowded environment and monitoring.
Summary of the invention
The technical scheme that the present invention solves the problems of the technologies described above is as follows:
Based on an abnormal behaviour discrimination method for kinematic constraint, comprise training process, training process comprises the following steps:
(1), input first group of video, pre-service carried out to input video two field picture, comprise adopt that quick calibrating method carries out demarcating, denoising etc.;
(2), to calibrated input video two field picture carry out background modeling, then from two field picture, extract Optical-flow Feature by optical flow method;
(3), with the kinematic constraint of pedestrian, the Optical-flow Feature in prospect is analyzed and chosen;
(4), with SVM selected Optical-flow Feature is learnt, obtain the SVM classifier that can be used for abnormal behaviour identification.
(5) parameter of SVM classifier model, is write down.
On the basis of technique scheme, the present invention can also do following improvement.
It is further, described that what from two field picture, extract that Optical-flow Feature takes is Lucas-Kanade algorithm.
Further, described kinematic constraint model is set up by Bayesian formula, and formula is:
P(u,v|I(x,y,t))＝αP(u,v)ΠP(I(x
i,y
i,t
i)|u,v)
Wherein, P (u, v) is prior probability, represent the two dimensional motion speed probability of object, α is the direct proportion factor, P (u, v|I (x, y, t)) be space (x, its two dimensional motion speed probability under gray-scale value condition when y) being in moment t, P (I (x, y, t) | u, the probability of the grey scale change being v) object when moving, i.e. posterior probability.
Further, analyzing with the kinematic constraint of pedestrian the Optical-flow Feature in prospect and choosing in described step 3, specifically comprises the following steps:
(1), from the foreground image after background detection, from the image-region of pedestrian, Optical-flow Feature is selected;
(2), go out to meet the Optical-flow Feature of threshold requirement by kinematic constraint model discrimination, described threshold value is the threshold value of posterior probability P (I (x, y, t) | u, v) being carried out to Select to use.
Further, the scope of described α between zero and one,
Further, described threshold value is optimized in the iterative process of SVM modeling process, obtains optimal value.
Further, the present invention also comprises testing process, and the step of described testing process is:
(1), input second group of video, pre-service carried out to input video two field picture, comprise adopt that quick calibrating method carries out demarcating, denoising etc.;
(2), to calibrated input video two field picture carry out background modeling, then from two field picture, extract Optical-flow Feature by optical flow method;
(3), with the kinematic constraint of pedestrian, the threshold value obtained according to the training stage, analyzes the Optical-flow Feature in prospect and chooses;
(4), by the SVM classifier that training process obtains judge selected Optical-flow Feature, if to abnormal behaviour (having known that this input belongs to abnormal behaviour in advance) correct judgment, then this sorter provable is reliable.
The invention has the beneficial effects as follows: the present invention is combined with kinematic constraint by being chosen by Optical-flow Feature, the SVM classifier that may be used for abnormal behaviour and judge is obtained at training process, the known second group of video of behavior classification in testing process by input, test training the model obtained, the correctness of testing model and reliability, also can play the effect extensive further to model simultaneously.Pedestrian movement's constraint is used for optimizing Optical-flow Feature and selects by the present invention first, effectively eliminate noise light stream characteristic sum and enhance interested Optical-flow Feature (flowfeaturesofinterest), and accurate judgement can be made to the abnormal behaviour such as to run.
Accompanying drawing explanation
Fig. 1 is the inventive method schematic diagram;
Fig. 2 is the schematic diagram utilizing Bayes Modeling;
Fig. 3 have employed the Optical-flow Feature result of calculation of kinematic constraint and does not adopt the result of calculation comparison diagram of Optical-flow Feature of kinematic constraint.
Embodiment
Be described principle of the present invention and feature below in conjunction with accompanying drawing, example, only for explaining the present invention, is not intended to limit scope of the present invention.
Based on an abnormal behaviour identification system for kinematic constraint, comprise pretreatment module, background modeling module, light stream extraction module, flow field selection module, kinematic constraint module and SVM classifier module.
Described pretreatment module is used for carrying out pre-service to input video two field picture, comprises demarcation, denoising etc.;
Described background modeling module is used for carrying out background modeling to pretreated input video two field picture;
Described light stream extraction module is used for extracting Optical-flow Feature two field picture after background modeling;
Described kinematic constraint module is used for the kinematic constraint with pedestrian, filters out qualified Optical-flow Feature according to threshold value,
Described SVM classifier module is used for learning selected Optical-flow Feature.In the training stage, by iteration and the Optimization Solution of SVM algorithm, obtain optimum model parameter and threshold value; At test phase, test known sample with the model that training obtains, whether the accuracy of detection model and reliability, can make the behavior of pedestrian (comprising abnormal behaviour) and judging accurately.
Embodiment 1
As shown in Figure 1, a kind of abnormal behaviour discrimination method based on kinematic constraint, comprises training process and testing process;
Described training process comprises the following steps:
(1), input first group of video, pre-service carried out to input video two field picture, comprise adopt that quick calibrating method carries out demarcating, denoising etc.;
(2), to calibrated input video two field picture carry out background modeling, then from two field picture, extract Optical-flow Feature by optical flow method;
(3), with the kinematic constraint of pedestrian, according to certain threshold value, the Optical-flow Feature in prospect is analyzed and chosen;
(4), with SVM selected Optical-flow Feature is learnt, obtain the SVM classifier that can be used for abnormal behaviour identification.
(5) parameter of SVM classifier model, is write down.
1, Optical-flow Feature definition
Light stream is widely used in Computer Vision Task, and as recognition of face, gait modeling, object tracking etc., two kinds of optical flow algorithms used are at present Horn-Schunck algorithm and Lucas-Kanade algorithm.Horn-Schunck algorithm has more smooth light stream, global information and precise time differential, but relatively slow and have rough boundary profile.The boundary error of Lucas-Kanade algorithm is more, but algorithm more simply can calculate fast.Consider computing velocity, what we adopted is Lucas-Kanade algorithm.
When not considering kinematic constraint, obtain Optical-flow Feature step and comprise optical flow computation and light stream extraction step.The Optical-flow Feature vector now obtained is the vector of the movable information implying unique point in two field picture, the priori not containing any pedestrian movement's constraint.
The generation of Optical-flow Feature vector is divided into two steps:
The first step, finds the proper vector point of human body image block, the head of such as certain pedestrian and the angle summit of left shoulder.
Second step, by the proper vector point of the previous frame image corresponding to found proper vector point, the head of such as same pedestrian and the angle summit of left shoulder.
2, Optical-flow Feature extracts
All light streams in selected foreground image can write P=[p (1), p (2), ..., p (i)], i=1..N, wherein p (i)=[Lx, Ly, Mx, My, α], the number of light stream is N (can calculate from optical flow algorithm), Lx and Ly represents abscissa value and the ordinate value of an optical flow position in image block respectively, Mx and My represents abscissa value and the ordinate value of corresponding light stream amplitude (being equivalent to velocity amplitude) respectively, and α represents deflection angle value.Because the light stream number in each image block is different, so also need to carry out standardization to each image block, the Optical-flow Feature after process can write Q=[p (x
1), p (x
2) ..., p (x
i)], wherein p (x
i) elect from p (i).Input parameter finally can use formula (2.1) to describe:
(2.1)I＝[p(x
1),p(x
2),...,p(x
i)],i＝1..N
Have selected Optical-flow Feature as the input of sorter after, problem can be summed up as the classification problem of two classes or multiclass.Because SVM is used as the sorter of pedestrian movement's behavior, selected input parameter is Optical-flow Feature position in image block, direction and speed etc.
3, the introducing of kinematic constraint
When not considering kinematic constraint, the Optical-flow Feature of pedestrian behavior is by the direct input as SVM training classifier.Although Optical-flow Feature can play a role behavior identification, because the robustness of optical flow method under complex environment is not high, need to introduce kinematic constraint.
Find after the Optical-flow Feature of pedestrian under congested conditions is analyzed, the motor pattern of pedestrian meets the sports rule with relating to persons, with other moving object as compared with vehicle travelled etc., there is obvious difference (motion of car is a kind of rigid motion, and movement velocity is all different with pedestrian with track); And the motor pattern of pedestrian, the velocity characteristic of particularly walking under normal circumstances, be have very much regular.Bayesian criterion is used to come pedestrian movement's constraint modeling, as shown in Figure 2.
Modeling process has two steps:
The first step, guarantee only from the image-region of pedestrian, to select Optical-flow Feature (in crowded environment, be difficult to judge whether a pixel is positioned among the image-region of people, a kind of solution of this problem is the algorithm adopting a kind of robust and detect people fast, or from the foreground image after background detection selected element, the latter is the method that this method adopts), the ratio that the Optical-flow Feature chosen accounts for Optical-flow Feature sum is P (A);
Second step, the Optical-flow Feature of threshold requirement is gone out to meet by kinematic constraint model discrimination, to make the calculating of Optical-flow Feature more accurate, thus judge whether selected pixel meets human motion constraint, the ratio that the Optical-flow Feature filtered out accounts for the Optical-flow Feature before screening is P (B), in the training process, P (A) and P (B) can reach a balance dynamically;
In second step, this method adopts bayes method (Bayesian) to come kinematic constraint modeling.The object that bayes method is used for human motion analysis calculates the posterior probability in Given Graph picture.Posterior probability is that the two dimensional motion speed (u, v) by calculating gray scale I (x, y, t) and object that t is positioned at (x, y) place obtains.
(3.1)P(u,v|I(x,y,t))＝αP(u,v)P(I(x,y,t)|u,v)
In formula (3.1), if hypothesis image is that what to observe from different positions is continuously independently with the time, then have:
(3.2)P(u,v|I(x,y,t))＝αP(u,v)ΠP(I(x
i,y
i,t
i)|u,v)
Wherein, P (u, v) is prior probability, represent the two dimensional motion speed probability of object, α is the direct proportion factor, P (u, v|I (x, y, t) be) its two dimensional motion speed probability under the gray-scale value condition of space (x, y) when being in moment t, can obtain in optical flow computation, P (I (x, y, t) | u, v) probability of the grey scale change that is object when moving, i.e. posterior probability, then according to threshold value to P (I (x, y, t) | u, v) carry out filtering out qualified Optical-flow Feature, threshold value is arbitrary value in the starting stage;
Go out to meet the Optical-flow Feature of threshold requirement by kinematic constraint model discrimination after, selected Optical-flow Feature as SVM algorithm input and learn, in the iterative of SVM algorithm, threshold value and SVM model parameter can constantly be optimized, final threshold value and the SVM model parameter obtaining optimum.
Method of the present invention also comprises testing process, and described testing process comprises the following steps:
(1), input second group of video, pre-service carried out to input video two field picture, comprise adopt that quick calibrating method carries out demarcating, denoising etc.;
(2), to calibrated input video two field picture carry out background modeling, then from two field picture, extract Optical-flow Feature by optical flow method;
(3), with the kinematic constraint of pedestrian, the threshold value obtained according to the training stage, analyzes the Optical-flow Feature in prospect and chooses;
(4) SVM classifier, by training process obtained judges selected Optical-flow Feature, if to abnormal behaviour (having known that this input belongs to abnormal behaviour in advance) correct judgment, then this sorter provable is reliable.
4, experimental result
Kinematic constraint is used for the result that under crowded environment, Optical-flow Feature calculates and a class abnormal behaviour is monitored by this experiment shows.In experiment, the resolution of monitor video is 320 × 240, and average length of time is 30 minutes.For every section of image of 30 minutes, wherein, within first 10 minutes, for model training, latter 20 minutes for model measurement.
Consider the Optical-flow Feature result of calculation of kinematic constraint.
Fig. 3 describes the Optical-flow Feature result of calculation that one have employed kinematic constraint, and left figure is the result of calculation not adopting kinematic constraint, and right figure is the result of calculation adopting kinematic constraint.As can see from Figure 3, kinematic constraint optimizes Optical-flow Feature and calculates, and eliminate noise Optical-flow Feature, pedestrian's Optical-flow Feature is with it more remarkable.As can also be seen from Figure, the kinematic constraint of car and the kinematic constraint of pedestrian have obvious difference simultaneously.
The foregoing is only better case study on implementation of the present invention, not in order to limit the present invention, within the spirit and principles in the present invention all, any amendment done, equivalent replacement, improvement etc., all should be included within protection scope of the present invention.
Claims (7)
1. based on an abnormal behaviour discrimination method for kinematic constraint, it is characterized in that, comprise training process, training process comprises the following steps:
(1), input first group of video, adopt quick calibrating method to demarcate to two field picture;
(2), to calibrated input video two field picture carry out background modeling, then from two field picture, extract Optical-flow Feature by optical flow method;
(3), with the kinematic constraint of pedestrian, the Optical-flow Feature in prospect is analyzed and chosen;
(4), with SVM selected Optical-flow Feature is learnt, obtain the SVM classifier that can be used for abnormal behaviour identification;
(5) parameter of SVM classifier model, is write down.
2. utilize the abnormal behaviour discrimination method based on kinematic constraint described in claim 1, it is characterized in that, described what from two field picture, extract that Optical-flow Feature takes is Lucas-Kanade algorithm.
3. utilize the abnormal behaviour discrimination method based on kinematic constraint described in claim 1, it is characterized in that, described kinematic constraint model is set up by Bayesian formula, and formula is:
P(u,v|I(x,y,t))＝αP(u
,v)ΠP(I(x
i,y
i,t
i)|u,v)
Wherein, P (u, v) is prior probability, represent the two dimensional motion speed probability of object, α is the direct proportion factor, P (u, v|I (x, y, t)) be space (x, its two dimensional motion speed probability under gray-scale value condition when y) being in moment t, P (I (x, y, t) | u, the probability of the grey scale change being v) object when moving, i.e. posterior probability.
4. utilize the abnormal behaviour discrimination method based on kinematic constraint described in claim 3, it is characterized in that, analyzing with the kinematic constraint of pedestrian the Optical-flow Feature in prospect and choosing in described step (3), specifically comprises the following steps:
(1), from the foreground image after background detection, from the image-region of pedestrian, Optical-flow Feature is selected;
(2), go out to meet the Optical-flow Feature of threshold requirement by kinematic constraint model discrimination, described threshold value is the threshold value of posterior probability P (I (x, y, t) | u, v) being carried out to Select to use.
5. utilize the abnormal behaviour discrimination method based on kinematic constraint described in claim 3, it is characterized in that, the scope of described α between zero and one.
6. utilize the abnormal behaviour discrimination method based on kinematic constraint described in claim 4, it is characterized in that, described threshold value is optimized in the iterative process of SVM modeling process, obtains optimal value.
7. utilize the abnormal behaviour discrimination method based on kinematic constraint described in claim 4, it is characterized in that, also comprise testing process, the step of described testing process is:
(1), input second group of video, adopt quick calibrating method to demarcate to input video two field picture;
(2), to calibrated input video two field picture carry out background modeling, from two field picture, extract Optical-flow Feature by optical flow method simultaneously;
(3), with the kinematic constraint of pedestrian, the threshold value obtained according to the training stage, analyzes the Optical-flow Feature in prospect and chooses;
(4) the SVM model parameter input SVM model, by training process obtained, judges selected Optical-flow Feature with SVM model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510551661.8A CN105046285B (en) | 2015-08-31 | 2015-08-31 | A kind of abnormal behaviour discrimination method based on kinematic constraint |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510551661.8A CN105046285B (en) | 2015-08-31 | 2015-08-31 | A kind of abnormal behaviour discrimination method based on kinematic constraint |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105046285A true CN105046285A (en) | 2015-11-11 |
| CN105046285B CN105046285B (en) | 2018-08-17 |
Family
ID=54452814
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510551661.8A Active CN105046285B (en) | 2015-08-31 | 2015-08-31 | A kind of abnormal behaviour discrimination method based on kinematic constraint |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105046285B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469054A (en) * | 2015-11-25 | 2016-04-06 | 天津光电高斯通信工程技术股份有限公司 | Model construction method of normal behaviors and detection method of abnormal behaviors |
| CN110222616A (en) * | 2019-05-28 | 2019-09-10 | 浙江大华技术股份有限公司 | Pedestrian's anomaly detection method, image processing apparatus and storage device |
| CN111046797A (en) * | 2019-12-12 | 2020-04-21 | 天地伟业技术有限公司 | Oil pipeline warning method based on personnel and vehicle behavior analysis |
| CN113762027A (en) * | 2021-03-15 | 2021-12-07 | 北京京东振世信息技术有限公司 | Abnormal behavior identification method, device, equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101271527A (en) * | 2008-02-25 | 2008-09-24 | 北京理工大学 | Exception action detecting method based on athletic ground partial statistics characteristic analysis |
| US20120148097A1 (en) * | 2010-12-14 | 2012-06-14 | Electronics And Telecommunications Research Institute | 3d motion recognition method and apparatus |
| CN102663429A (en) * | 2012-04-11 | 2012-09-12 | 上海交通大学 | Method for motion pattern classification and action recognition of moving target |
-
2015
- 2015-08-31 CN CN201510551661.8A patent/CN105046285B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101271527A (en) * | 2008-02-25 | 2008-09-24 | 北京理工大学 | Exception action detecting method based on athletic ground partial statistics characteristic analysis |
| US20120148097A1 (en) * | 2010-12-14 | 2012-06-14 | Electronics And Telecommunications Research Institute | 3d motion recognition method and apparatus |
| CN102663429A (en) * | 2012-04-11 | 2012-09-12 | 上海交通大学 | Method for motion pattern classification and action recognition of moving target |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105469054A (en) * | 2015-11-25 | 2016-04-06 | 天津光电高斯通信工程技术股份有限公司 | Model construction method of normal behaviors and detection method of abnormal behaviors |
| CN105469054B (en) * | 2015-11-25 | 2019-05-07 | 天津光电高斯通信工程技术股份有限公司 | The model building method of normal behaviour and the detection method of abnormal behaviour |
| CN110222616A (en) * | 2019-05-28 | 2019-09-10 | 浙江大华技术股份有限公司 | Pedestrian's anomaly detection method, image processing apparatus and storage device |
| CN110222616B (en) * | 2019-05-28 | 2021-08-31 | 浙江大华技术股份有限公司 | Pedestrian abnormal behavior detection method, image processing device and storage device |
| CN111046797A (en) * | 2019-12-12 | 2020-04-21 | 天地伟业技术有限公司 | Oil pipeline warning method based on personnel and vehicle behavior analysis |
| CN113762027A (en) * | 2021-03-15 | 2021-12-07 | 北京京东振世信息技术有限公司 | Abnormal behavior identification method, device, equipment and storage medium |
| CN113762027B (en) * | 2021-03-15 | 2023-09-08 | 北京京东振世信息技术有限公司 | Abnormal behavior identification method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105046285B (en) | 2018-08-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Min et al. | A new approach to track multiple vehicles with the combination of robust detection and two classifiers | |
| Hu et al. | Data-driven estimation of driver attention using calibration-free eye gaze and scene features | |
| CN112750150B (en) | Vehicle flow statistical method based on vehicle detection and multi-target tracking | |
| Basharat et al. | Learning object motion patterns for anomaly detection and improved object detection | |
| CN108647649B (en) | Method for detecting abnormal behaviors in video | |
| CN103258332B (en) | A kind of detection method of the moving target of resisting illumination variation | |
| CN104281853A (en) | Behavior identification method based on 3D convolution neural network | |
| CN102609720B (en) | Pedestrian detection method based on position correction model | |
| CN104574439A (en) | Kalman filtering and TLD (tracking-learning-detection) algorithm integrated target tracking method | |
| CN104318263A (en) | Real-time high-precision people stream counting method | |
| CN104992453A (en) | Target tracking method under complicated background based on extreme learning machine | |
| CN102831378B (en) | The detection and tracking method and system of people | |
| CN110298297A (en) | Flame identification method and device | |
| CN105046285A (en) | Abnormal behavior identification method based on motion constraints | |
| CN103488993A (en) | Crowd abnormal behavior identification method based on FAST | |
| Šilar et al. | The obstacle detection on the railway crossing based on optical flow and clustering | |
| CN107085703A (en) | Merge face detection and the automobile passenger method of counting of tracking | |
| CN109166137A (en) | For shake Moving Object in Video Sequences detection algorithm | |
| Khan et al. | Estimating speeds of pedestrians in real-world using computer vision | |
| Khan | Estimating Speeds and Directions of Pedestrians in Real-Time Videos: A solution to Road-Safety Problem. | |
| Amjad et al. | Kalman filter-based convolutional neural network for robust tracking of froth-middling interface in a primary separation vessel in presence of occlusions | |
| Nguyen et al. | Lane detection and tracking based on fully convolutional networks and probabilistic graphical models | |
| Tang et al. | Salient moving object detection using stochastic approach filtering | |
| Bhaskar et al. | Enhanced and effective parallel optical flow method for vehicle detection and tracking | |
| Fu et al. | Real-time adaptive multi-classifier multi-resolution visual tracking framework for unmanned aerial vehicles |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20200709 Address after: 350015 No. 99 Denglong Road, Fuzhou Economic and Technological Development Zone, Fujian Province Patentee after: YANGO University Address before: 430000, No. 31, Wuhuan Avenue, Dongxihu District, Hubei, Wuhan Patentee before: WUHAN YINGSHI INTELLIGENT SCIENCE & TECHNOLOGY Co.,Ltd. |
|
| TR01 | Transfer of patent right |