CN103428407B - A kind of method for detecting fought in video - Google Patents
A kind of method for detecting fought in video Download PDFInfo
- Publication number
- CN103428407B CN103428407B CN201210164916.1A CN201210164916A CN103428407B CN 103428407 B CN103428407 B CN 103428407B CN 201210164916 A CN201210164916 A CN 201210164916A CN 103428407 B CN103428407 B CN 103428407B
- Authority
- CN
- China
- Prior art keywords
- block
- strenuous exercise
- frame
- certain
- image
- 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
Abstract
The invention provides the method for detecting fought in a kind of video for the detection of fighting being suitable under complicated light condition, comprise the following steps：(1) validity feature point value is obtained：Carry out feature point detection, characteristic point estimation and quantization；(2) strenuous exercise's block is obtained：By image block, every piece of feature direction histogram is counted, histogram entropy is calculated, if histogram entropy is taken as strenuous exercise's block more than certain threshold value；(3) strenuous exercise's fraction is added up：When certain frame block is strenuous exercise's block, its strenuous exercise's fraction adds certain value；If it is not, this block strenuous exercise fraction just is reduced into certain proportion；(4) strenuous exercise's connected region is exported：If the violent confidence values of certain block exceed given threshold, strenuous exercise's block is taken as, and strenuous exercise's block of connection is constituted into connected region, if connected region area is more than given threshold, is just exported as strenuous exercise's connected region.
Description
Technical field
The present invention relates to a kind of video detecting method, the method for detecting fought in especially a kind of video.
Background technology
In recent years, Video Supervision Technique is increasingly used in safety-security area, with the continuous growth of monitoring data amount,
So that monitoring personnel needs to use up the substantial amounts of time to data progress artificial filter, useful video information is therefrom selected, and it is numerous
Work that is multiple, repeating usually makes them unable to do what one wishes.Monitoring personnel is lighter in the urgent need to there is one kind to work together, it is easier to grasp
Make and the product of management substitutes original system.Then, intelligent video technology is arisen at the historic moment, into the life of people.
Intelligent video is that computer vision methods are incorporated into intelligent monitoring.This technology is included by sequence of video images
Automatically moved or detection, target classification and the behavior understanding of static target in terms of content, it is therefore an objective to image with
Mapping relations are set up between iamge description, so as to enable a computer to analyze and understand the content in video pictures.
Detection technique of fighting is the important technology in intelligent video analysis, in the monitoring of the place safeties such as square, hotel, prison
There are wide application scenarios in.But major part researcher is primarily focused in the more satisfactory scene of solution at present
Fight detection technique.However, for complicated light scene, conventional method is fought, and Detection results are not often good, and appearance is a large amount of to be reported by mistake.
If improving detection threshold value, more fail to report is caused again.
The content of the invention
The invention provides the method for detecting fought in a kind of video for the detection of fighting being suitable under complicated light condition.
The method for detecting fought in the video for realizing the object of the invention, comprises the following steps：
(1) validity feature point value is obtained：Carry out feature point detection, characteristic point estimation and quantization；
(2) strenuous exercise's block is obtained：By image block, every piece of feature direction histogram is counted, histogram entropy is calculated, such as
Fruit histogram entropy is taken as strenuous exercise's block more than certain threshold value；
(3) strenuous exercise's fraction is added up：When certain frame block is strenuous exercise's block, its strenuous exercise's fraction adds certain value；
If it is not, this block strenuous exercise fraction just is reduced into certain proportion；
(4) strenuous exercise's connected region is exported：If the violent confidence values of certain block exceed given threshold, strenuous exercise is taken as
Block, and strenuous exercise's block of connection is constituted into connected region, if connected region area is more than given threshold, just as acutely fortune
Dynamic connected region output.
The method for detecting fought in the video of the present invention has the beneficial effect that：
The method for detecting fought in the video of the present invention, improves the accuracy for detection of fighting, reduces rate of false alarm, especially
It is that effect is good under complicated light conditions.
Brief description of the drawings
The flow chart for the method for detecting that Fig. 1 fights in the video for the present invention.
Embodiment
As shown in figure 1, the method for detecting fought in the video of the present invention, comprises the following steps：
(1) validity feature point value is obtained：Carry out feature point detection, characteristic point estimation and quantization；
(2) strenuous exercise's block is obtained：By image block, every piece of feature direction histogram is counted, histogram entropy is calculated, such as
Fruit histogram entropy is taken as strenuous exercise's block more than certain threshold value；
(3) strenuous exercise's fraction is added up：When certain frame block is strenuous exercise's block, its strenuous exercise's fraction adds certain value；
If it is not, this block strenuous exercise fraction just is reduced into certain proportion；
(4) strenuous exercise's connected region is exported：If the violent confidence values of certain block exceed given threshold, strenuous exercise is taken as
Block, and strenuous exercise's block of connection is constituted into connected region, if connected region area is more than given threshold, just as acutely fortune
Dynamic connected region output.
Validity feature point value obtaining step is as follows：
The feature point detection unit, sequentially reads in each two field picture；For nth frame image, the Harr i s of image are calculated
Angle point, N is the integer more than or equal to 1；
The characteristic point motion estimation unit, to nth frame image, calculates N-1 two field pictures and this frame image features point pair
The characteristic point answered, composition characteristic point pair, and using characteristic point to forming motion arrow with terminating point as the starting of motion vector
Amount, N is the integer more than 1；
Motion vector magnitude, is less than the characteristic point of certain threshold value to filter by the feature spot moving direction quantifying unit first
Remove；Again by remaining point to according to the number that angular quantification is 0- (A-1)；
Strenuous exercise's block obtaining step is as follows：
The statistic histogram unit, the KxK of the size block such as divides an image into first, counts in this image (N
Frame) before N1 frames (taking N-N1+1, N-N1+2 ..., N frame) this block in all directions characteristic point number, form a length of A's
Histogram, if that histogram is respectively worth and less than certain threshold value, handles next image block；
It is described to ask for entropy unit, it is right by the histogram normalization of this image this block (i.e. divided by sum for being respectively worth of histogram)
Normalized value asks for entropy；
Strenuous exercise's block identifying unit, when entropy exceedes certain threshold value, it is strenuous exercise's block to determine that this block of this frame；
Strenuous exercise's connected region output step is as follows：
The continuous acutely motion frame number resets unit, and the continuous acutely motion frame number of all image blocks of the first two field picture is set
It is zero；
The continuous acutely motion frame number summing elements, to a frame input picture, to each image block, if it is acutely transported
Dynamic fraction is more than certain threshold value, and the continuous acutely motion frame number of this block of this frame just is set into the upper continuous frame number of this block of frame plus one, otherwise will
This frame this block frame number is set to zero.
The stable connected region output unit, to a frame input picture, the continuous violent motion frame number that will abut against exceedes
The block composition connected region of certain threshold value, when the area of this connected region is more than certain threshold value, just output alarm, and by this connected region
It is used as the alarm region for detection of fighting.
The above-described embodiments are merely illustrative of preferred embodiments of the present invention, not to the model of the present invention
Enclose and be defined, under the premise of design spirit of the present invention is not departed from, this area ordinary skill technical staff is to the technology of the present invention side
In various modifications and improvement that case is made, the protection domain that claims of the present invention determination all should be fallen into.
Claims (1)
1. the method for detecting fought in a kind of video, comprises the following steps：
Step 1 validity feature point value is obtained：Carry out feature point detection, characteristic point estimation and quantization；
Feature point detection unit, sequentially reads in each two field picture；For nth frame image, the Harris angle points of image are calculated, N is big
In the integer equal to 1；
Characteristic point motion estimation unit, to nth frame image, calculates N-1 two field pictures feature corresponding with this frame image features point
Point, composition characteristic point pair, and using characteristic point to forming motion vector with terminating point as the starting of motion vector, N is big
In 1 integer；
Feature spot moving direction quantifying unit, is less than the characteristic point of certain first threshold to filtering out by motion vector magnitude first；
Again by remaining point to according to the number that angular quantification is 0~(A-1)；
Step 2 strenuous exercise block is obtained：By image block, every piece of feature direction histogram is counted, histogram entropy is calculated, if
Histogram entropy is taken as strenuous exercise's block more than certain 3rd threshold value；
Statistic histogram unit, the KxK of the size block such as divides an image into first, and statistics includes before this nth frame image
N1 frames including nth frame, that is, take the number of all directions characteristic point in N-N1+1, N-N1+2 ..., N frame, N1 ＜ N, this block, is formed
A length of A histogram, if that histogram is respectively worth and less than certain Second Threshold, handles next image block；
Divided by the sum that is respectively worth of histogram entropy unit is asked for, the histogram of this image this block is normalized, i.e., to normalized value
Ask for entropy；
Strenuous exercise's block identifying unit, when entropy exceedes certain the 3rd threshold value, it is strenuous exercise's block to determine that this block of this frame；
Step 3 adds up strenuous exercise's fraction：When certain frame block is strenuous exercise's block, its strenuous exercise's fraction adds certain value；Such as
Fruit is not that this block strenuous exercise fraction just is reduced into certain proportion；
Step 4 strenuous exercise connected region output step is as follows：
Continuous acutely motion frame number resets unit, and the continuous acutely motion frame number of all image blocks of the first two field picture is set into zero；
Continuous acutely motion frame number summing elements, to a frame input picture, to each image block, if its strenuous exercise's fraction is big
In certain the 4th threshold value, the continuous acutely motion frame number of this block of this frame is just set to the upper continuous frame number of this block of frame and plus one, otherwise by this frame
This block frame number is set to zero；
Stable connected region output unit, to a frame input picture, the continuous violent motion frame number that will abut against exceedes certain the 5th threshold
The block composition connected region of value, when the area of this connected region is more than certain the 5th threshold value, just output alarm, and by this connected region
It is used as the alarm region for detection of fighting.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210164916.1A CN103428407B (en) | 2012-05-25 | 2012-05-25 | A kind of method for detecting fought in video |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210164916.1A CN103428407B (en) | 2012-05-25 | 2012-05-25 | A kind of method for detecting fought in video |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103428407A CN103428407A (en) | 2013-12-04 |
| CN103428407B true CN103428407B (en) | 2017-08-25 |
Family
ID=49652541
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210164916.1A Expired - Fee Related CN103428407B (en) | 2012-05-25 | 2012-05-25 | A kind of method for detecting fought in video |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103428407B (en) |
Families Citing this family (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104092993A (en) * | 2014-07-15 | 2014-10-08 | 广州市番禺奥莱照明电器有限公司 | Street lamp controlling and security monitoring device, system and method based on video analysis |
| US10121062B2 (en) * | 2014-11-03 | 2018-11-06 | Koninklijke Philips N.V. | Device, system and method for automated detection of orientation and/or location of a person |
| CN110298323B (en) * | 2019-07-02 | 2021-10-15 | 中国科学院自动化研究所 | Frame-fighting detection method, system and device based on video analysis |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101161201A (en) * | 2007-11-08 | 2008-04-16 | 珠海友通科技有限公司 | Method for registrating external circumstance DSA elasticity automatically |
| CN101216941A (en) * | 2008-01-17 | 2008-07-09 | 上海交通大学 | Motion estimation method under violent illumination variation based on corner matching and optic flow method |
| CN101336856A (en) * | 2008-08-08 | 2009-01-07 | 西安电子科技大学 | Information acquisition and transfer method of auxiliary vision system |
| 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 |
| CN102231792A (en) * | 2011-06-29 | 2011-11-02 | 南京大学 | Electronic image stabilization method based on characteristic coupling |
| CN102427505A (en) * | 2011-09-29 | 2012-04-25 | 深圳市万兴软件有限公司 | Video image stabilization method and system on the basis of Harris Corner |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100194851A1 (en) * | 2009-02-03 | 2010-08-05 | Aricent Inc. | Panorama image stitching |
| CN101883258B (en) * | 2009-05-08 | 2011-11-16 | 上海弘视通信技术有限公司 | Violent crime detection system and detection method thereof |
| CN102360422A (en) * | 2011-10-19 | 2012-02-22 | 湖南德顺电子科技有限公司 | Violent behavior detecting method based on video analysis |
-
2012
- 2012-05-25 CN CN201210164916.1A patent/CN103428407B/en not_active Expired - Fee Related
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101161201A (en) * | 2007-11-08 | 2008-04-16 | 珠海友通科技有限公司 | Method for registrating external circumstance DSA elasticity automatically |
| CN101216941A (en) * | 2008-01-17 | 2008-07-09 | 上海交通大学 | Motion estimation method under violent illumination variation based on corner matching and optic flow method |
| CN101336856A (en) * | 2008-08-08 | 2009-01-07 | 西安电子科技大学 | Information acquisition and transfer method of auxiliary vision system |
| 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 |
| CN102231792A (en) * | 2011-06-29 | 2011-11-02 | 南京大学 | Electronic image stabilization method based on characteristic coupling |
| CN102427505A (en) * | 2011-09-29 | 2012-04-25 | 深圳市万兴软件有限公司 | Video image stabilization method and system on the basis of Harris Corner |
Non-Patent Citations (1)
| Title |
|---|
| 基于区域光流特征的异常行为检测;杜鉴豪等;《浙江大学学报(工学版)》;20110730;第45卷(第7期);第1162页左栏第24行至第1165页左栏第4行 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103428407A (en) | 2013-12-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104966304B (en) | Multi-target detection tracking based on Kalman filtering and nonparametric background model | |
| CN103903273B (en) | PM2.5 grade fast-evaluating system based on mobile phone terminal | |
| JP2006507775A (en) | Method and apparatus for measuring the quality of a compressed video sequence without criteria | |
| JP2008243181A (en) | Smoke detecting device and method thereof | |
| CN105141923B (en) | A kind of video concentration method and device | |
| CN103428407B (en) | A kind of method for detecting fought in video | |
| CN107564018A (en) | It is a kind of to utilize the method for improving iterative algorithm extraction target image | |
| CN106548142A (en) | Crowd's incident detection and appraisal procedure in a kind of video based on comentropy | |
| JP4926602B2 (en) | Smoke detector | |
| CN109727202A (en) | Gas infrared video Enhancement Method, device and storage medium | |
| CN103093197A (en) | Monitoring method and system for recognizing hanging behavior | |
| CN105930814A (en) | Method for detecting personnel abnormal gathering behavior on the basis of video monitoring platform | |
| CN103425958A (en) | Method for detecting non-movable objects in video | |
| CN106485703B (en) | Fuzzy detection method based on image gradient dct transform | |
| CN113781384A (en) | Video quality evaluation method and device | |
| CN115170894B (en) | Method and device for detecting smoke and fire | |
| CN104185022A (en) | Full-reference video quality evaluation method based on visual information distortion decomposition | |
| KR102177494B1 (en) | method of identifying abnormal-motion objects in compressed video by use of trajectory and pattern of motion vectors | |
| JP5119740B2 (en) | VIDEO SIGNAL PROCESSING DEVICE, VIDEO SIGNAL PROCESSING METHOD, VIDEO SIGNAL PROCESSING METHOD PROGRAM, AND RECORDING MEDIUM CONTAINING VIDEO SIGNAL PROCESSING METHOD PROGRAM | |
| Tang et al. | Canny edge detection codec using VLib on Davinci series DSP | |
| CN104754181B (en) | A kind of video noise estimation method and device | |
| CN115514929A (en) | Two-stage series storage system based on video compression data | |
| CN113810695A (en) | Video encoding method, apparatus and computer-readable storage medium | |
| KR102178952B1 (en) | method of providing object classification for compressed video by use of syntax-based MRPN-CNN | |
| CN113066077A (en) | Flame detection method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Address after: 100089 Haidian District, Beijing information industry base on the 3rd Street, building two, floor C, No. 228, paragraph Applicant after: Frame robot technology (Beijing) Co., Ltd. Address before: 100089 Haidian District, Beijing information industry base on the 3rd Street, building two, floor C, No. 228, paragraph Applicant before: Xinzheng Electronic Technology (Beijing) Co., Ltd. |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170825 Termination date: 20190525 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |