CN104732236B - A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping - Google Patents
A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping Download PDFInfo
- Publication number
- CN104732236B CN104732236B CN201510126186.XA CN201510126186A CN104732236B CN 104732236 B CN104732236 B CN 104732236B CN 201510126186 A CN201510126186 A CN 201510126186A CN 104732236 B CN104732236 B CN 104732236B
- Authority
- CN
- China
- Prior art keywords
- crowd
- movement
- area
- abnormal behaviour
- state
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Analysis (AREA)
Abstract
A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping.It includes crowd state anticipation and two processes of crowd's unusual checking, and wherein crowd state anticipation process is prejudged to crowd state using crowd state variability index R；Crowd's unusual checking process is to utilize crowd movement's ENERGY E simultaneouslynWith crowd movement direction entropy HnCarry out crowd's abnormal behaviour judgement.Crowd's abnormal behaviour intelligent detecting method provided by the invention based on layered shaping not only has good detection result to crowd's abnormal behaviour, and the process of learning training is not needed, flexibility simultaneously is strong, operand is low, it can ensure real-time well, being suitable for square, subway, railway station, airport etc., the crowd is dense and the higher place of public safety demand.
Description
Technical field
The invention belongs to intelligent Video Surveillance Technology fields, more particularly to a kind of crowd's exception row based on layered shaping
For intelligent detecting method.
Technical background
In recent years, as the Mass disturbance to endanger public security frequently occurs, demand of the people to public safety is more next
It is higher, it is had been to be concerned by more and more people with visual analysis so analyzing crowd's abnormal behaviour under crowd's scene.Wherein people
Group's behavioural analysis is the most challenging problem of computer vision field.
Chinese Patent No. 201110090467.6 discloses a kind of based on the crowd's abnormal behaviour for improving social force model
Detection method, but this method is only applicable to highdensity crowd's scene.Chinese Patent No. 201210403819.3,
201310437367.5 the crowd's anomaly detection method proposed in 201310494769.9 all refers to the mistake of learning training
Journey, and result is more coarse, and algorithm complexity is high, and operand is big, and therefore, it is difficult to ensure real-time.
Invention content
To solve the above-mentioned problems, crowd's abnormal behaviour intelligence based on layered shaping that the purpose of the present invention is to provide a kind of
It can detection method.
In order to achieve the above object, crowd's abnormal behaviour intelligent detecting method packet provided by the invention based on layered shaping
Crowd state anticipation and two processes of crowd's unusual checking are included, wherein：Crowd state anticipation process includes carrying out in order
The following steps：
1) crowd's foreground is extracted using mixed Gauss model to video input；
2) median filter process is carried out using median filter to crowd's foreground that said extracted arrives, then calculates crowd
Foreground area area S；
4) difference is taken to adjacent two frames video and obtains frame difference region area S2；
4) using the crowd foreground area area S obtained in step 2) as foreground mask operator to the frame that is obtained in step 3)
Poor region area S2Mask operation is done, crowd's moving area area S is thus obtained1；
5) according to the crowd's moving area area obtained in the crowd foreground area area S obtained in step 2) and step 4)
S1Crowd state variability index R is calculated, wherein
6) crowd state is prejudged using crowd state variability index R, works as R<R0When, illustrate that crowd is in normal shape
State, R0Indicate crowd state threshold value；Work as R>R0When, illustrate that crowd movement's state changes, there may be unusual condition, into
One step is detected by following steps；
Crowd's unusual checking process includes the following steps carried out in order：
A Harris Corner Detections) are carried out to video input；
B L-K optical flow trackings) are carried out to the above-mentioned Harris angle points detected, obtain crowd movement's vector field wi=u,
V }, wherein wiRepresent the motion vector field of ith feature point in n-th frame；
C) according to step B) in obtained crowd movement's vector field, calculate the speed of each characteristic point,
D) according to step C) the obtained speed of characteristic point, further calculate crowd movement's ENERGY E of n-th framen,Wherein miRepresent the quality of the ith feature point of n-th frame, it is assumed here that the quality m of all characteristic pointsi
=1, n-th frame shares j characteristic point；
E) according to step B) in obtained crowd movement's vector field, further obtain crowd movement's directional spreding histogram:h
(i)={ ki,0<i<=8 }, wherein h (i) represents the number for the motion vector that each direction includes, and is divided into 8 directions in total；
F) according to step E) in obtained crowd movement's directional spreding histogram obtain crowd's directional spreding probability P (i),0<i<=8；Wherein m is the sum of crowd's motion vector,
G) according to step F) in obtained crowd's directional spreding probability, further calculate crowd movement direction entropy Hn,
H) according to step D) in obtained crowd movement's ENERGY EnWith step G) in obtained crowd movement direction entropy HnIt carries out
Crowd's abnormal behaviour judges, if En>E0And Hn>H0, judge that abnormal behaviour, wherein H has occurred in the people group0Indicate crowd movement direction
The threshold value of entropy.
Crowd's abnormal behaviour intelligent detecting method provided by the invention based on layered shaping is not only to crowd's abnormal behaviour
With good detection result, and the process of learning training is not needed, while flexibility is strong, operand is low, can be well
Ensure real-time, being suitable for square, subway, railway station, airport etc., the crowd is dense and the higher place of public safety demand.
Description of the drawings
Fig. 1 is crowd's abnormal behaviour intelligent detecting method overall flow figure provided by the invention based on layered shaping.
Fig. 2 is that crowd state prejudges in crowd's abnormal behaviour intelligent detecting method provided by the invention based on layered shaping
Process flow diagram flow chart.
Fig. 3 is crowd's abnormal behaviour in crowd's abnormal behaviour intelligent detecting method provided by the invention based on layered shaping
Detection process flow chart.
Fig. 4 is that crowd direction divides schematic diagram.
Fig. 5 is crowd's direction of motion distribution histogram.
Specific implementation mode
Crowd's abnormal behaviour intelligence based on layered shaping to provided by the invention in the following with reference to the drawings and specific embodiments
Detection method is described in detail.
As shown in Fig. 1-Fig. 3, crowd's abnormal behaviour intelligent detecting method provided by the invention based on layered shaping includes
Crowd state prejudges and two processes of crowd's unusual checking, wherein：It includes carrying out in order that crowd state, which prejudges process,
The following steps：
1) crowd's foreground is extracted using mixed Gauss model to video input；
By to a large amount of demographic data collection test find, when crowd is normal, crowd's moving area be maintained at it is one smaller and
Stable level, when crowd is abnormal behavior, crowd's moving area increases rapidly.Crowd's moving area and crowd's foreground area
Difference, the latter describe the region for being not belonging to background, the former describes the region that crowd moves in foreground area.
2) median filter process is carried out using median filter to crowd's foreground that said extracted arrives, then calculates crowd
Foreground area area S；
5) difference is taken to adjacent two frames video and obtains frame difference region area S2；
4) using the crowd foreground area area S obtained in step 2) as foreground mask operator to the frame that is obtained in step 3)
Poor region area S2Mask operation is done, crowd's moving area area S is thus obtained1；
5) according to the crowd's moving area area obtained in the crowd foreground area area S obtained in step 2) and step 4)
S1Crowd state variability index R is calculated, wherein
6) crowd state is prejudged using crowd state variability index R, works as R<R0When, illustrate that crowd is in normal shape
State, R0Indicate crowd state threshold value, value size can be by obtaining a large amount of test result statistical analyses；Work as R>R0When, illustrate people
Group's motion state changes, and may have unusual condition, is further detected by following steps；
Crowd's unusual checking process includes the following steps carried out in order：
A Harris Corner Detections) are carried out to video input；
B L-K optical flow trackings) are carried out to the above-mentioned Harris angle points detected, obtain crowd movement's vector field wi=u,
V }, wherein wiRepresent the motion vector field of ith feature point in n-th frame；
C) according to step B) in obtained crowd movement's vector field, calculate the speed of each characteristic point,
D) according to step C) the obtained speed of characteristic point, further calculate crowd movement's ENERGY E of n-th framen,Wherein miRepresent the quality of the ith feature point of n-th frame, it is assumed here that the quality m of all characteristic pointsi
=1, n-th frame shares j characteristic point；
E) according to step B) in obtained crowd movement's vector field, further obtain crowd movement's directional spreding histogram:h
(i)={ ki,0<i<=8 }, wherein h (i) represents the number for the motion vector that each direction includes, and is divided into 8 directions in total；
Crowd movement's directional spreding rule is one of important feature of crowd state.Crowd movement direction is single with 45 degree
Position, is divided into 8 directions, as shown in Figure 4.Crowd movement's directional spreding histogram is as shown in Figure 5.
F) according to step E) in obtained crowd movement's directional spreding histogram obtain crowd's directional spreding probability P (i),0<i<=8；Wherein m is the sum of crowd's motion vector,
G) according to step F) in obtained crowd's directional spreding probability, further calculate crowd movement direction entropy Hn,
H) according to step D) in obtained crowd movement's ENERGY EnWith step G) in obtained crowd movement direction entropy HnIt carries out
Crowd's abnormal behaviour judges, if En>E0And Hn>H0, judge that abnormal behaviour, wherein H has occurred in the people group0Indicate crowd movement direction
The threshold value of entropy, value size can be by obtaining to a large amount of test result is for statistical analysis.
Claims (1)
1. a kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping, it is characterised in that：It includes that crowd state is pre-
Sentence with two processes of crowd's unusual checking, wherein：It includes the following steps carried out in order that crowd state, which prejudges process,：
1) crowd's foreground is extracted using mixed Gauss model to video input；
2) median filter process is carried out using median filter to crowd's foreground that said extracted arrives, then calculates crowd's foreground
Region area S；
3) difference is taken to adjacent two frames video and obtains frame difference region area S2；
4) using the crowd foreground area area S obtained in step 2) as foreground mask operator to the areas Zheng Cha that are obtained in step 3)
Domain area S2Mask operation is done, crowd's moving area area S is thus obtained1；
5) according to the crowd's moving area area S obtained in the crowd foreground area area S obtained in step 2) and step 4)1Meter
Crowd state variability index R is calculated, wherein
6) crowd state is prejudged using crowd state variability index R, works as R<R0When, illustrate that crowd is in normal condition, R0
Indicate crowd state threshold value；Work as R>R0When, illustrate that crowd movement's state changes, there may be unusual condition, further
It is detected by following steps；
Crowd's unusual checking process includes the following steps carried out in order：
A Harris Corner Detections) are carried out to video input；
B L-K optical flow trackings) are carried out to the above-mentioned Harris angle points detected, obtain crowd movement's vector field wi={ u, v }, wherein
wiRepresent the motion vector field of ith feature point in n-th frame；
C) according to step B) in obtained crowd movement's vector field, calculate the speed of each characteristic point,
D) according to step C) the obtained speed of characteristic point, further calculate crowd movement's ENERGY E of n-th framen,
Wherein miRepresent the quality of the ith feature point of n-th frame, it is assumed here that the quality m of all characteristic pointsi=1, n-th frame shares j
A characteristic point；
E) according to step B) in obtained crowd movement's vector field, further obtain crowd movement's directional spreding histogram:h(i)
={ ki,0<i<=8 }, wherein h (i) represents the number for the motion vector that each direction includes, and is divided into 8 directions in total；
F) according to step E) in obtained crowd movement's directional spreding histogram obtain crowd's directional spreding probability P (i),Wherein m is the sum of crowd's motion vector,
G) according to step F) in obtained crowd's directional spreding probability, further calculate crowd movement direction entropy Hn,
H) according to step D) in obtained crowd movement's ENERGY EnWith step G) in obtained crowd movement direction entropy HnCarry out crowd
Abnormal behaviour judges, if En>E0And Hn>H0, judge that abnormal behaviour, wherein H has occurred in the people group0Indicate crowd movement direction entropy
Threshold value.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510126186.XA CN104732236B (en) | 2015-03-23 | 2015-03-23 | A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510126186.XA CN104732236B (en) | 2015-03-23 | 2015-03-23 | A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104732236A CN104732236A (en) | 2015-06-24 |
| CN104732236B true CN104732236B (en) | 2018-08-24 |
Family
ID=53456110
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510126186.XA Expired - Fee Related CN104732236B (en) | 2015-03-23 | 2015-03-23 | A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104732236B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241950A (en) * | 2018-10-19 | 2019-01-18 | 杭州电子科技大学 | A kind of crowd panic state identification method based on enthalpy Distribution Entropy |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103693532B (en) * | 2013-12-26 | 2016-04-27 | 江南大学 | Act of violence detection method in a kind of lift car |
| CN106446922B (en) * | 2015-07-31 | 2019-10-22 | 中国科学院大学 | A kind of crowd's abnormal behaviour analysis method |
| 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 |
| CN106203357A (en) * | 2016-07-11 | 2016-12-07 | 浙江宇视科技有限公司 | The detection method of a kind of gathering of people and device |
| CN108596028B (en) * | 2018-03-19 | 2022-02-08 | 昆明理工大学 | Abnormal behavior detection algorithm based on video recording |
| CN109299700A (en) * | 2018-10-15 | 2019-02-01 | 南京地铁集团有限公司 | Subway group abnormality behavioral value method based on crowd density analysis |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751678A (en) * | 2009-12-16 | 2010-06-23 | 北京智安邦科技有限公司 | Method and device for detecting violent crowd movement |
| 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 |
| KR20130103213A (en) * | 2012-03-09 | 2013-09-23 | 고려대학교 산학협력단 | Detection and analysis of abnormal crowd behavior in h.264 compression domain |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7558404B2 (en) * | 2005-11-28 | 2009-07-07 | Honeywell International Inc. | Detection of abnormal crowd behavior |
-
2015
- 2015-03-23 CN CN201510126186.XA patent/CN104732236B/en not_active Expired - Fee Related
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101751678A (en) * | 2009-12-16 | 2010-06-23 | 北京智安邦科技有限公司 | Method and device for detecting violent crowd movement |
| 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 |
| KR20130103213A (en) * | 2012-03-09 | 2013-09-23 | 고려대학교 산학협력단 | Detection and analysis of abnormal crowd behavior in h.264 compression domain |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109241950A (en) * | 2018-10-19 | 2019-01-18 | 杭州电子科技大学 | A kind of crowd panic state identification method based on enthalpy Distribution Entropy |
| CN109241950B (en) * | 2018-10-19 | 2021-11-02 | 杭州电子科技大学 | Crowd panic state identification method based on enthalpy distribution entropy |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104732236A (en) | 2015-06-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104732236B (en) | A kind of crowd's abnormal behaviour intelligent detecting method based on layered shaping | |
| CN103824070B (en) | A kind of rapid pedestrian detection method based on computer vision | |
| CN102542289B (en) | Pedestrian volume statistical method based on plurality of Gaussian counting models | |
| CN113011367B (en) | Abnormal behavior analysis method based on target track | |
| CN103839065A (en) | Extraction method for dynamic crowd gathering characteristics | |
| CN103903008B (en) | A kind of method and system of the mist grade based on image recognition transmission line of electricity | |
| CN103246896B (en) | A kind of real-time detection and tracking method of robustness vehicle | |
| CN102521565A (en) | Garment identification method and system for low-resolution video | |
| CN105678811A (en) | Motion-detection-based human body abnormal behavior detection method | |
| CN107274432B (en) | A kind of intelligent video monitoring method | |
| CN105303191A (en) | Method and apparatus for counting pedestrians in foresight monitoring scene | |
| CN105138982A (en) | Crowd abnormity detection and evaluation method based on multi-characteristic cluster and classification | |
| CN105404847A (en) | Real-time detection method for object left behind | |
| CN103488993A (en) | Crowd abnormal behavior identification method based on FAST | |
| CN109918971A (en) | Number detection method and device in monitor video | |
| CN103426179A (en) | Target tracking method and system based on mean shift multi-feature fusion | |
| CN105096342A (en) | Intrusion detection algorithm based on Fourier descriptor and histogram of oriented gradient | |
| Su et al. | A new local-main-gradient-orientation HOG and contour differences based algorithm for object classification | |
| Jiang et al. | An Approach for Crowd Density and Crowd Size Estimation. | |
| CN103106796A (en) | Vehicle detection method and device of intelligent traffic surveillance and control system | |
| CN109299700A (en) | Subway group abnormality behavioral value method based on crowd density analysis | |
| CN106127813B (en) | The monitor video motion segments dividing method of view-based access control model energy sensing | |
| CN104091351B (en) | Number method of counting based on clustering procedure | |
| CN101877135B (en) | Moving target detecting method based on background reconstruction | |
| CN107240111B (en) | Edge communication segmentation passenger flow statistical method |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180824 Termination date: 20190323 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |