CN112364680B - Abnormal behavior detection method based on optical flow algorithm - Google Patents
Abnormal behavior detection method based on optical flow algorithm Download PDFInfo
- Publication number
- CN112364680B CN112364680B CN202010985851.1A CN202010985851A CN112364680B CN 112364680 B CN112364680 B CN 112364680B CN 202010985851 A CN202010985851 A CN 202010985851A CN 112364680 B CN112364680 B CN 112364680B
- Authority
- CN
- China
- Prior art keywords
- optical flow
- amplitude
- entropy
- histogram
- abnormal behavior
- 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.)
- Active
Links
- 230000003287 optical effect Effects 0.000 title claims abstract description 101
- 206010000117 Abnormal behaviour Diseases 0.000 title claims abstract description 51
- 238000001514 detection method Methods 0.000 title claims abstract description 21
- 230000000739 chaotic effect Effects 0.000 claims abstract description 7
- 230000005856 abnormality Effects 0.000 claims abstract description 5
- 230000006399 behavior Effects 0.000 claims description 27
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000000034 method Methods 0.000 claims description 8
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 4
- 208000012661 Dyskinesia Diseases 0.000 description 2
- AYFVYJQAPQTCCC-GBXIJSLDSA-N L-threonine Chemical compound C[C@@H](O)[C@H](N)C(O)=O AYFVYJQAPQTCCC-GBXIJSLDSA-N 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 238000002474 experimental method Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000012544 monitoring process Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 239000000126 substance Substances 0.000 description 2
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 238000005452 bending Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 230000006378 damage Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 208000014674 injury Diseases 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses an abnormal behavior detection method based on an optical flow algorithm, which is implemented according to the following steps: step 1, extracting optical flow information by an optical flow algorithm: extracting optical flow information generated when a human body moves by adopting a Farnesback dense optical flow algorithm; step 2, extracting behavior characteristics: counting the optical flow information extracted in the step 1 into a direction amplitude histogram; step 3, result analysis and abnormality determination: whether abnormal behavior occurs is judged by calculating the direction and the amplitude entropy of the histogram, and the larger the direction and the amplitude entropy of the histogram are, the more chaotic the current motion is, the greater the possibility of occurrence of the abnormal behavior is indicated, and the problem of misjudgment of the abnormal behavior easily existing in the prior art is solved.
Description
Technical Field
The invention belongs to the technical field of intelligent video monitoring, and relates to an abnormal behavior detection method based on an optical flow algorithm.
Background
At present, public places such as parking lots, railway stations and the like define abnormal behaviors as follows: the cradle and robbery behaviors which are different from the normal walking behaviors of the human body and have the characteristics of high movement speed and disordered movement direction are defined as abnormal behaviors.
In recent years, rapid behaviors with illegal crime properties such as frame taking, robbery and the like frequently occur in many public places, such as parking lots, railway stations, markets and the like, which seriously disturb social public order and bring people with much personal injury and property loss. In order to minimize the hazards, video surveillance research is also gaining attention. Numerous cameras are installed in various public places of parking lots and railway stations, and safety of society is guaranteed. If the intelligent monitoring system alarms in time under the condition of detecting abnormality in some densely populated public places, the life and property safety of masses can be effectively ensured.
The abnormal behavior of the human body is generally represented by abnormal movement speed and abnormal movement direction, and the traditional detection method is characterized in that the characteristic extraction method has the problems of poor detection effect, easy erroneous judgment of the abnormal behavior, complex calculation method and the like.
Disclosure of Invention
The invention aims to provide an abnormal behavior detection method based on an optical flow algorithm, which solves the problem of misjudgment of abnormal behavior in the prior art.
The technical scheme adopted by the invention is that the abnormal behavior detection method based on the optical flow algorithm is implemented according to the following steps:
step 1, extracting optical flow information by an optical flow algorithm: extracting optical flow information generated when a human body moves by adopting a Farnesback dense optical flow algorithm;
step 2, extracting behavior characteristics: counting the optical flow information extracted in the step 1 into a direction amplitude histogram;
step 3, result analysis and abnormality determination: whether abnormal behavior occurs is judged by calculating the direction and the amplitude entropy of the histogram, and the larger the direction and the amplitude entropy of the histogram is, the more chaotic the current motion is, and the greater the possibility of occurrence of the abnormal behavior is indicated.
The invention is also characterized in that:
the step 2 is specifically implemented according to the following steps:
step 2.1: conversion of planar rectangular coordinates to polar coordinates
The inter-frame optical flow field calculated by the farnebback dense optical flow algorithm is an original feature obtained from the video image, and one optical flow vector can be represented by a four-dimensional vector (x, y, u, v), wherein (x, y) represents the spatial position information of the optical flow vector in the image, and (u, v) represents the magnitudes of components of the optical flow vector in the horizontal direction and the vertical direction, respectively;
step 2.2: converting vector information of optical flow into histogram by Matlab software
Describing the behavior of human body during movement by adopting a direction amplitude histogram, wherein the abscissa is set into a plurality of zones, and the size of each zone isThe ordinate represents the statistics of the magnitude of the amplitude over each interval.
In step 2.1, the optical flow vector component (u, v) is converted from a plane rectangular coordinate form to a (r, θ) form under polar coordinates, and the optical flow size can be obtained through calculation and analysisDirection of optical flow->
In step 2.2, the abscissa is set to 12 intervals.
The step 3 is specifically implemented according to the following steps:
step 3.1: let the amplitude of the ith interval be h i The sum of the directional times of the i section is s i Before calculating the entropy of the direction and the amplitude, the probability of the direction and the amplitude of each section is calculated firstly, and the calculation formula is as follows:
S dir : sum of the number of times of direction of the direction amplitude histogram;
S hyp : sum of the magnitudes of the directional magnitude histogram;
the directional entropy and the amplitude entropy are calculated as follows:
directional entropy:
amplitude entropy:
the directional entropy and the amplitude entropy are one of important factors for judging whether abnormal behaviors exist, so that the directional entropy and the amplitude entropy can be integrated into a formula, and the formula is expressed as follows:
W＝E S ×E h ；
the larger W indicates the more chaotic the current motion, indicating a greater likelihood of abnormal behavior.
The beneficial effects of the invention are as follows: the abnormal behavior detection method based on the optical flow algorithm solves the problems of poor detection effect, easy misjudgment of abnormal behavior, complex calculation method and the like in the prior art; and extracting optical flow information by adopting a Farnesback dense optical flow algorithm, analyzing the behavior of the optical flow information by using the optical flow information generated during human body movement, and judging whether the optical flow information belongs to abnormal behaviors or not. The optical flow method does not need to acquire the background of the image in advance, and the calculation result only depends on the relative motion of continuous frames and is not influenced by complex environments; the concept of entropy is introduced, which is used in thermodynamics to describe the state of a substance, and its physical meaning is to represent the degree of confusion of a system. When abnormal behaviors occur, the characteristics of disordered directions of all parts of the body and the like exist, so that whether the abnormal behaviors exist can be judged by adopting whether the entropy value is larger than a certain threshold value.
Drawings
FIG. 1 is a flow chart of an abnormal behavior detection method based on an optical flow algorithm of the present invention;
FIG. 2 is a technical roadmap of an abnormal behavior detection method based on an optical flow algorithm of the invention;
FIG. 3a is a normal walking W curve of an abnormal behavior detection method based on an optical flow algorithm;
FIG. 3b is a drawing behavior W curve of the abnormal behavior detection method based on the optical flow algorithm;
fig. 3c is a robbery behavior W curve of an abnormal behavior detection method based on an optical flow algorithm of the present invention.
Detailed Description
The invention will be described in detail below with reference to the drawings and the detailed description.
The invention discloses an abnormal behavior detection method based on an optical flow algorithm, which is implemented according to the following steps as shown in fig. 1:
step 1, extracting optical flow information by an optical flow algorithm: extracting optical flow information generated when a human body moves by adopting a Farnesback dense optical flow algorithm;
as shown in fig. 2, specifically, in OpenCV, the API of calcopticalflow farnebback () is used to perform computation of dense optical flow, and parameters of the calcopticalflow farnebback function are specifically set to (prvs, next, none,0.5,3,15,3,5,1.2,0), the first parameter represents input of a previous frame image, the second parameter represents input of a subsequent frame image, the third parameter represents output optical flow, the fourth parameter represents pyramid scaling parameter, the fifth parameter represents pyramid layer number, the sixth parameter represents window size, the seventh parameter represents iteration number, the eighth parameter represents pixel neighborhood size, the ninth parameter represents gaussian standard deviation, and the tenth parameter flag represents computation mode.
Step 2, extracting behavior characteristics: counting the optical flow information extracted in the step 1 into a direction amplitude histogram;
the step 2 is specifically implemented according to the following steps:
step 2.1: conversion of planar rectangular coordinates to polar coordinates
The inter-frame optical flow field calculated by the Farnesback dense optical flow algorithm is the original feature derived from the video image, and an optical flow vector can be usedExpressed by a four-dimensional vector (x, y, u, v), wherein (x, y) represents the spatial position information of the optical flow vector in the image, and (u, v) represents the magnitudes of the components of the optical flow vector in the horizontal direction and the vertical direction, respectively; in step 2.1, the optical flow vector component (u, v) is converted from a plane rectangular coordinate form to a (r, θ) form under polar coordinates, and the optical flow size can be obtained through calculation and analysisDirection of optical flow->
Step 2.2: converting vector information of optical flow into histogram by Matlab software
The conventional direction histogram is a non-parametric estimation method, and although it can represent the optical flow characteristics of a moving object, the change of the amplitude in the direction in the optical flow information is ignored, so that the direction amplitude histogram is used to describe the behavior of the human body during the movement, the abscissa is set as a plurality of intervals, and the size of each interval isThe ordinate represents the statistics of the magnitude of the amplitude over each interval. In step 2.2, the abscissa is set to 12 intervals.
Step 3, result analysis and abnormality determination: whether abnormal behavior occurs is judged by calculating the direction and the amplitude entropy of the histogram, and the larger the direction and the amplitude entropy of the histogram is, the more chaotic the current motion is, and the greater the possibility of occurrence of the abnormal behavior is indicated.
The step 3 is specifically implemented according to the following steps:
step 3.1: let the amplitude of the ith interval be h i The sum of the directional times of the i section is s i Before calculating the entropy of the direction and the amplitude, the probability of the direction and the amplitude of each section is calculated firstly, and the calculation formula is as follows:
S dir : sum of the number of times of direction of the direction amplitude histogram;
S hyp : sum of the magnitudes of the directional magnitude histogram;
the directional entropy and the amplitude entropy are calculated as follows:
directional entropy:
amplitude entropy:
the directional entropy and the amplitude entropy are one of important factors for judging whether abnormal behaviors exist, so that the directional entropy and the amplitude entropy can be integrated into a formula, and the formula is expressed as follows:
W＝E S ×E h ；
the larger W indicates the more chaotic the current motion, indicating a greater likelihood of abnormal behavior.
The invention relates to an abnormal behavior detection method based on an optical flow algorithm, which comprises the following steps: the function of step 1 is to extract the optical flow features in the images of the video sequence.
The optical flow characteristics are extracted by adopting a Farnesback dense optical flow algorithm, and the principle is as follows: it is first assumed that the brightness between adjacent frames is constant, and the frame taking time of adjacent video frames is continuous, or the motion of the object between adjacent frames is relatively small, while maintaining spatial consistency, i.e. the pixels of the same sub-image have the same motion. The main idea of farnebback dense optical flow is to approximate the neighborhood information of each pixel with a polynomial, e.g. consider the quadratic polynomial f (x) =x T Ax+b T x+c, A is a symmetric matrix, b is a vector, c is a scalar, f (x) is expressed as an approximate description of the neighborhood information of the pixel, A is obtained by least squares weighted fitting of the neighborhood information of the pixel, and the weight coefficients are related to the pixel size and position of the neighborhood. For images of the previous frameRepresenting the displacement of two frames of images with d, then +.>Since the appearance information of the pixels in the image scene is unchanged during the inter-frame motion, we can get the same corresponding coefficients, in case of a non-singular matrix, the displacement + ->And then tracking the characteristic points of the image by combining the optimization and adjustment of the error and the image pyramid.
The optical flow characteristics are extracted by adopting the Farnesback dense optical flow algorithm, and the method has the advantages that: the dense optical flow algorithm can calculate the offset of all points on the image to form a dense optical flow field, and pixel-level image registration can be performed by using the dense optical flow field, so that the effect after registration is obviously better than that of sparse optical flow registration.
Examples
1) In this example, a CASIA behavior data set is used, which has a total of 1446 video data, and is captured by cameras distributed in three different perspectives in an outdoor parking lot environment. The data are divided into single-person behaviors and multi-person interaction behaviors, wherein the single-person behaviors comprise the following eight different behaviors: "walking", "running", "bending over", "jumping", "squatting", "tipping", "loitering" and "car crashing", each of which has 24 persons taking about 4 shots per person. The multi-person interaction behavior includes seven different behaviors: robbery, racking, trailing, catch up, bump, meet, and overrun, once or twice for every two persons. The background in the video is complex, the light change is obvious, the frame rate is 25fps, the resolution is 320 multiplied by 240, huffyuv coding compression is adopted, and the video is stored in the form of an avi file.
In the experiment, the cradling and robbing behaviors different from the normal behaviors (walking) of the human body and having the characteristics of high movement speed and inconsistent movement directions are defined as abnormal behaviors.
2) Experimental training used the Windows 10 system, the processor inter (R) Core (TM) i5-8250U CPU@1.60GHz 1.80GHz,8.00GB memory, and the integrated development environment was Pycharm, openCV open source library and Matlab R2016a.
3) In order to extract the optical flow information in the motion area more accurately, an effective optical flow is introduced on the basis of the traditional optical flow, namely, only the optical flow points which have larger contribution to abnormal behavior detection are reserved through screening of the optical flow amplitude, and the rest optical flow points are omitted.
The effective optical flow field can be obtained from the following formula:
wherein thre: the threshold for screening the optical flow point is selected after a plurality of experiments, and the method adopts thre=0.3.
4) Compared with normal behavior, when the cradling and robbing behavior occurs, the motion direction is disordered, the motion is violent, the value of W is increased, the change of the optical flow characteristic can be well reflected by the value of W, and W change curves of different behaviors are shown in fig. 3 (a) - (c).
(a) The W change curves of the normal walking, the cradling and the robbing in one scene are respectively, the W average values of the three different behaviors are compared, the W average value of the normal walking is smaller, the W average value of the cradling and the robbing is larger, and therefore, the W value has better distinction for the normal behavior and the abnormal behavior, and the possibility of abnormal behavior is larger when the W value is larger.
The abnormal behavior detection method based on the optical flow algorithm solves the problems of poor detection effect, easy misjudgment of abnormal behavior, complex calculation method and the like in the prior art; and extracting optical flow information by adopting a Farnesback dense optical flow algorithm, analyzing the behavior of the optical flow information by using the optical flow information generated during human body movement, and judging whether the optical flow information belongs to abnormal behaviors or not. The optical flow method does not need to acquire the background of the image in advance, and the calculation result only depends on the relative motion of continuous frames and is not influenced by complex environments; the concept of entropy is introduced, which is used in thermodynamics to describe the state of a substance, and its physical meaning is to represent the degree of confusion of a system. When abnormal behaviors occur, the characteristics of disordered directions of all parts of the body and the like exist, so that whether the abnormal behaviors exist can be judged by adopting whether the entropy value is larger than a certain threshold value.
Claims (4)
1. The abnormal behavior detection method based on the optical flow algorithm is characterized by comprising the following steps:
step 1, extracting optical flow information by an optical flow algorithm: extracting optical flow information generated when a human body moves by adopting a Farnesback dense optical flow algorithm;
step 2, extracting behavior characteristics: counting the optical flow information extracted in the step 1 into a direction amplitude histogram;
step 3, result analysis and abnormality determination: judging whether abnormal behaviors occur or not by calculating the direction and amplitude entropy of the histogram, wherein the larger the direction and amplitude entropy of the histogram is, the more chaotic the current motion is, and the greater the possibility of occurrence of the abnormal behaviors is indicated:
step 3.1: let the amplitude of the ith interval be h i The sum of the directional times of the i section is s i Before calculating the entropy of the direction and the amplitude, the probability of the direction and the amplitude of each section is calculated firstly, and the calculation formula is as follows:
S dir : sum of the number of times of direction of the direction amplitude histogram;
S hyp : sum of the magnitudes of the directional magnitude histogram;
the directional entropy and the amplitude entropy are calculated as follows:
directional entropy:
amplitude entropy:
the directional entropy and the amplitude entropy are one of important factors for judging whether abnormal behaviors exist, so that the directional entropy and the amplitude entropy can be integrated into a formula, and the formula is expressed as follows:
W＝E S ×E h ；
the larger W indicates the more chaotic the current motion, indicating a greater likelihood of abnormal behavior.
2. The method for detecting abnormal behavior based on the optical flow algorithm according to claim 1, wherein the step 2 is specifically implemented according to the following steps:
step 2.1: conversion of planar rectangular coordinates to polar coordinates
The inter-frame optical flow field calculated by the farnebback dense optical flow algorithm is an original feature obtained from the video image, and one optical flow vector can be represented by a four-dimensional vector (x, y, u, v), wherein (x, y) represents the spatial position information of the optical flow vector in the image, and (u, v) represents the magnitudes of components of the optical flow vector in the horizontal direction and the vertical direction, respectively;
step 2.2: converting vector information of optical flow into histogram by Matlab software
Describing the behavior of human body during movement by adopting a direction amplitude histogram, wherein the abscissa is set into a plurality of zones, and the size of each zone isThe ordinate represents the statistics of the magnitude of the amplitude over each interval.
3. The method for detecting abnormal behavior based on optical flow algorithm according to claim 2, wherein in step 2.1, optical flow is performedThe vector component (u, v) is converted from a plane rectangular coordinate form to a (r, theta) form under polar coordinates, and the magnitude of the optical flow can be obtained through calculation and analysisDirection of optical flow->
4. The method for detecting abnormal behavior based on optical flow algorithm according to claim 2, wherein in the step 2.2, the abscissa is set to 12 intervals.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010985851.1A CN112364680B (en) | 2020-09-18 | 2020-09-18 | Abnormal behavior detection method based on optical flow algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010985851.1A CN112364680B (en) | 2020-09-18 | 2020-09-18 | Abnormal behavior detection method based on optical flow algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112364680A CN112364680A (en) | 2021-02-12 |
| CN112364680B true CN112364680B (en) | 2024-03-05 |
Family
ID=74516561
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010985851.1A Active CN112364680B (en) | 2020-09-18 | 2020-09-18 | Abnormal behavior detection method based on optical flow algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112364680B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113537165B (en) * | 2021-09-15 | 2021-12-07 | 湖南信达通信息技术有限公司 | Detection method and system for pedestrian alarm |
| CN113870323B (en) * | 2021-09-30 | 2024-04-02 | 西安交通大学 | Structure surface micro crack identification method based on optical flow information entropy |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102629384A (en) * | 2012-02-28 | 2012-08-08 | 成都三泰电子实业股份有限公司 | Method for detecting abnormal behavior during video monitoring |
| CN102880444A (en) * | 2012-08-24 | 2013-01-16 | 浙江捷尚视觉科技有限公司 | Fighting detecting method based on stereoscopic vision motion field analysis |
| CN103810718A (en) * | 2012-11-15 | 2014-05-21 | 浙江大华技术股份有限公司 | Method and device for detection of violently moving target |
| CN104123544A (en) * | 2014-07-23 | 2014-10-29 | 通号通信信息集团有限公司 | Video analysis based abnormal behavior detection method and system |
| CN106327791A (en) * | 2016-08-31 | 2017-01-11 | 朱明� | Exceptional event real-time alarm method and system |
| CN106326840A (en) * | 2016-08-11 | 2017-01-11 | 电子科技大学 | Method for detecting abnormal behaviors of group of people based on optical flow histogram entropy |
| CN107967440A (en) * | 2017-09-19 | 2018-04-27 | 北京工业大学 | A kind of monitor video method for detecting abnormality based on multizone mutative scale 3D-HOF |
Family Cites Families (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| DE102007008021A1 (en) * | 2007-02-15 | 2008-08-21 | Tecdata Ag | Method for measuring information |
| US8396268B2 (en) * | 2010-03-31 | 2013-03-12 | Isis Innovation Limited | System and method for image sequence processing |
| AU2015268735A1 (en) * | 2015-12-15 | 2017-06-29 | Canon Kabushiki Kaisha | Method, system and apparatus for spatially arranging a plurality of video frames for display |
| US10345046B2 (en) * | 2017-05-25 | 2019-07-09 | Northeastern University | Fault diagnosis device based on common information and special information of running video information for electric-arc furnace and method thereof |
-
2020
- 2020-09-18 CN CN202010985851.1A patent/CN112364680B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102629384A (en) * | 2012-02-28 | 2012-08-08 | 成都三泰电子实业股份有限公司 | Method for detecting abnormal behavior during video monitoring |
| CN102880444A (en) * | 2012-08-24 | 2013-01-16 | 浙江捷尚视觉科技有限公司 | Fighting detecting method based on stereoscopic vision motion field analysis |
| CN103810718A (en) * | 2012-11-15 | 2014-05-21 | 浙江大华技术股份有限公司 | Method and device for detection of violently moving target |
| CN104123544A (en) * | 2014-07-23 | 2014-10-29 | 通号通信信息集团有限公司 | Video analysis based abnormal behavior detection method and system |
| CN106326840A (en) * | 2016-08-11 | 2017-01-11 | 电子科技大学 | Method for detecting abnormal behaviors of group of people based on optical flow histogram entropy |
| CN106327791A (en) * | 2016-08-31 | 2017-01-11 | 朱明� | Exceptional event real-time alarm method and system |
| CN107967440A (en) * | 2017-09-19 | 2018-04-27 | 北京工业大学 | A kind of monitor video method for detecting abnormality based on multizone mutative scale 3D-HOF |
Non-Patent Citations (6)
| Title |
|---|
| Huifang Qian 等.Abnormal Behavior Detection and Recognition Method Based on Improved ResNet Model.《CMC-Computers, Materials & Continua》.2020,第65卷(第3期),2153-2167. * |
| Rensso Victor Hugo Mora Colque 等.Histograms of Optical Flow Orientation and Magnitude and Entropy to Detect Anomalous Events in Videos.《IEEE Transactions on Circuits and Systems for Video Technology》.2017,第27卷(第3期),376-682. * |
| 基于光流法的深度学习在工业运动检测的应用;周曼 等;《自动化与仪表》;第34卷(第7期);第1-2节 * |
| 基于监控视频的异常事件识别;丁茜 等;《光学仪器》;第41卷(第1期);第2-3节 * |
| 王燕妮 等.关键帧结合幅值直方图熵的异常行为检测算法.《计算机与数字工程》.2019,第47卷(第9期),2281-2285. * |
| 视频监控系统中异常行为检测软件设计与实现;安井然;《中国优秀硕士学位论文全文数据库 信息科技辑》;I138-1062 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112364680A (en) | 2021-02-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wen et al. | Detection, tracking, and counting meets drones in crowds: A benchmark | |
| WO2020173226A1 (en) | Spatial-temporal behavior detection method | |
| CN105844234B (en) | Method and equipment for counting people based on head and shoulder detection | |
| US20190130583A1 (en) | Still and slow object tracking in a hybrid video analytics system | |
| US20190130191A1 (en) | Bounding box smoothing for object tracking in a video analytics system | |
| Fuentes et al. | People tracking in surveillance applications | |
| WO2016173277A9 (en) | Video coding and decoding methods and apparatus | |
| CN106203513B (en) | A kind of statistical method based on pedestrian's head and shoulder multi-target detection and tracking | |
| CN111144247A (en) | Escalator passenger reverse-running detection method based on deep learning | |
| Zhu et al. | Video anomaly detection for smart surveillance | |
| KR20150065847A (en) | System and method for processing visual information for event detection | |
| CN112364680B (en) | Abnormal behavior detection method based on optical flow algorithm | |
| WO2008092393A1 (en) | Method of moving target tracking and number accounting | |
| Sengar et al. | Motion detection using block based bi-directional optical flow method | |
| Huang et al. | Motion detection with pyramid structure of background model for intelligent surveillance systems | |
| CN110782433A (en) | Dynamic information violent parabolic detection method and device based on time sequence and storage medium | |
| CN112380905B (en) | Abnormal behavior detection method based on histogram combination entropy of monitoring video | |
| Yu et al. | A robust method for detecting and counting people | |
| Cardile et al. | A vision-based system for elderly patients monitoring | |
| CN210072642U (en) | Crowd abnormal behavior detection system based on video monitoring | |
| Guo et al. | Pedestrian violence detection based on optical flow energy characteristics | |
| KR20200060868A (en) | multi-view monitoring system using object-oriented auto-tracking function | |
| Lin et al. | Compressed-domain fall incident detection for intelligent homecare | |
| CN112307895A (en) | Crowd gathering abnormal behavior detection method under community monitoring scene | |
| CN111915713A (en) | Three-dimensional dynamic scene creating method, computer equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |