CN102799935B - Human flow counting method based on video analysis technology - Google Patents
Human flow counting method based on video analysis technology Download PDFInfo
- Publication number
- CN102799935B CN102799935B CN201210208666.7A CN201210208666A CN102799935B CN 102799935 B CN102799935 B CN 102799935B CN 201210208666 A CN201210208666 A CN 201210208666A CN 102799935 B CN102799935 B CN 102799935B
- Authority
- CN
- China
- Prior art keywords
- people
- characteristic area
- frame
- feature
- target
- 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
Landscapes
- Image Analysis (AREA)
Abstract
The invention discloses a human flow counting method based on a video analysis technology, and relates to an intelligent video analysis technology. The counting method comprises the following steps of: (1) training a human head feature model; (2) acquiring a foreground point of a moving object by using an inter-frame difference; (3) performing human head feature extraction and identification; (4) performing human body local feature identification; (5) performing human head feature region tracking; and (6) performing human flow counting. The performance is stable, speed and efficiency are high, and the false alarm rate is low; and generality, portability and extensibility are high, and the method is suitable for equipment of various manufacturers and various intelligent video analysis systems.
Description
Technical field
The present invention relates to video intelligent analytical technology, particularly relate to and look a kind of the people flow rate statistical method based on Video Analysis Technology.
Background technology
In video intelligent analysis system, people flow rate statistical is a critical function of this system, it effectively can be applied to the people flow rate statistical of the public places such as market, bus, gateway, airport, subway station gateway, exhibition venue gateway, by gathering zones of different and Different periods flow of the people data, excavating, Comparison and analysis, for user management and decision-making provide important evidence.
Current existing people flow rate statistical method is mainly:
1, " the people flow rate statistical method and system based on intelligent video identification technology ", applicant: Xinlian-Weixun Science & Technology Development Co Ltd, Shanghai // Chinese patent, application number: 200810037799.Obtain the body templates in current frame video image according to the information scanning in standardized human body's template database, then two front line directions of body templates are judged, and then the flow of the people in statistics both direction.
2, " pedestrian traffic statistical method and system based on traffic monitoring facilities ", applicant: Beijing Jiaotong University // Chinese patent, application number: 201010155338.In detection block, carry out number of people detection according to standard number of people template, judge whether the number of people detected presses detection line, thus determine the effective number of people number in detection block.Then mate with previous testing result, and count according to the number of people of matching result to different motion direction.Finally according to the facilities of detection block and detection line, carry out traffic statistics.
Method 1 scans the body templates obtained in current frame video image, and body templates is applied in actual scene often exists more blocking; Method 2 mainly detects the number of people, between the number of people to block probability less, but detect that false-alarm also can correspondingly increase.
Summary of the invention
The object of the invention is to the shortcoming and defect overcoming prior art existence, a kind of people flow rate statistical method based on Video Analysis Technology is provided.
The object of the present invention is achieved like this:
The body local feature (people's head and shoulders combines) of frame-to-frame differences method and a kind of Corpus--based Method pattern recognition theory is adopted to detect recognition technology.Adopt frame-to-frame differences to obtain the foreground point information of moving object, number of people feature detection is carried out to the foreground point of motion, improve the speed of algorithm and reduce wrong report.Adopt body local feature (people's head and shoulders combines) inspection policies, can effectively eliminate the false-alarm detected, improve the accuracy rate of algorithm, be applicable to the scene that the various stream of people is more.
One, a kind of people flow rate statistical system based on Video Analysis Technology (abbreviation statistical system)
This statistical system comprises working environment: video monitoring platform, comprehensive access gate, intelligent management server;
Be provided with intellectual analysis server;
Its annexation is: video monitoring platform, comprehensive access gate, intelligent management server are connected successively with intellectual analysis server.
Principle of work
Intellectual analysis server is connected to intelligent management server, and intellectual analysis server is according to the IP(Internet protocol of intelligent management server) and port be connected to intelligent management server; When user asks video intelligent analysis task, this request is sent to intelligent management server, intelligent management server 30 records intellectual analysis server state, and by camera list equilibrium assignment to be detected to idle intellectual analysis server, intellectual analysis server taking turn equipment, obtain real-time video from camera and decode, obtain RGB (red, green, blue, RGB) data, then RGB data is analyzed, and testing result is reported to intelligent management server, result preserves by intelligent management server.User also can report to the police according to alarm type and date inquiries, statistics generating report forms.
Two, a kind of people flow rate statistical method based on Video Analysis Technology (abbreviation statistical method)
As Fig. 2, this statistical method comprises step:
1. number of people characteristic model-201 is trained
To gradient orientation histogram feature (hog) the sample training of the number of people, generate number of people characteristic model, namely generate number of people gradient orientation histogram feature samples model;
2. frame-to-frame differences obtains the foreground point-202 of moving object
Frame-to-frame differences method is adopted to obtain the foreground point information of moving object;
3. number of people feature extraction and identification-203
To the feature extraction algorithm of the image applications Corpus--based Method after frame-to-frame differences, the target in detected image, mates the target detected with sample pattern, tentatively determines the number of people characteristic area in this two field picture;
4. body local feature identification-204
To the number of people region decision detected, whether it exists shoulder feature, when there is shoulder feature, is then number of people characteristic area, otherwise is non-number of people characteristic area;
5. number of people characteristic area follows the tracks of-205
Adopt the tracking technique combined based on coupling and target following, the number of people detected is followed the tracks of, obtains the movement locus of pedestrian, and the frame number that number of people characteristic area occurs is added up;
6. flow of the people counting-206
Does when number of people target crosses line and region by assigned direction, and the frame number that number of people target occurs meet certain condition (what condition?) time, the counting on assigned direction is carried out to pedestrian, statistics the party flow of the people upwards.
The present invention has following advantages and good effect:
1, stable performance, speed is fast, and efficiency is high and rate of false alarm is low;
2, highly versatile, is applicable to the equipment of each producer;
3, portable strong, extendability is flexible;
4, all kinds of video intelligent analysis system is applicable to.
Accompanying drawing explanation
Fig. 1 is this statistical system block diagram, in figure
10-video monitoring platform,
11-the 1 video monitoring platform,
12-the 2nd video monitoring platform
1N-N video monitoring platform, N is natural number, N<10;
20-comprehensive access gate;
30-intelligent management server;
40-intellectual analysis server,
41-the 1st intellectual analysis server
4N-N intellectual analysis server, N is natural number, N<100.
Fig. 2 is this statistical method block diagram.
Fig. 3 sets up number of people sample pattern method flow diagram.
Fig. 4 is number of people object detection method process flow diagram.
Fig. 5 is people head's tracking detection method process flow diagram.
Embodiment
Describe in detail below in conjunction with drawings and Examples:
One, statistical system
1, overall
As Fig. 1, this statistical system comprises working environment: video monitoring platform 10, comprehensive access gate 20, intelligent management server 30;
Be provided with intellectual analysis server 40;
Its annexation is: video monitoring platform 10, comprehensive access gate 20, intelligent management server 30 are connected successively with intellectual analysis server 40.
2, functional part
1) video monitoring platform 10
For user provides the business such as remote collection, transmission, Storage and Processing of real-time audio and video and various alerting signal.
2) comprehensive access gate 20
Realize the statistics access of video monitoring platform.
3) intelligent management server 30
Realize intelligent resource management, be in charge of intellectual analysis resource.
4) intellectual analysis server 40
Intellectual analysis server 40 is functional entitys that video intelligent is analyzed, a corresponding station server in physical distribution.Intellectual analysis server 40 is by multiple VA(video analysis unit) form, each VA can the intellectual analysis of complete independently one road video.
Major function is:
1. video intelligent analytical algorithm is realized;
2. be linked into intelligent management server 30, managed concentratedly by intelligent management server 30;
3. receive the video intelligent analysis request of intelligent management server 30, obtain video from video monitoring platform 10 and analyze;
4. diagnostic result is reported intelligent management server 30.
A kind of people flow rate statistical method based on Video Analysis Technology of the present invention is implemented in the VA module of intellectual analysis server 40.
Specifically, the VA module of intellectual analysis server 40 comprises the functional software in general-purpose computer and implantation computer.
Two, statistical method
As Fig. 2, this statistical method is a kind of people flow rate statistical method based on Video Analysis Technology in summary of the invention.
1, number of people sample pattern method is set up
This method mainly adopts svm(to support/hold vector machine) sorter sets up number of people feature samples model, for people head's mark does not supply a model, relates to the step of total method 1..
As Fig. 3, performing step is as follows:
1. input number of people feature samples-301, input non-number of people feature samples-302;
2. hog feature extraction-303 is carried out to the sample of input, obtain the proper vector of sample;
3. apply svm sorter-304, the proper vector of input is trained;
4. by the computing of svm, the model-305 of number of people feature is obtained.
2, the detection method of number of people feature
This method mainly adopts frame-to-frame differences method to obtain the foreground information of moving object to video image, extract foreground point hog feature, and the target detected is mated with number of people feature samples masterplate, the number of people characteristic area that the match is successful is judged whether it exists shoulder feature, thus effectively distinguish number of people characteristic area and non-number of people characteristic area, relate to the step of total method 2. 3. 4..
As Fig. 4, performing step is as follows:
Hog and histogram of oriented gradient, it is the Feature Descriptor for target detection, the direction gradient number of times that image local occurs by this technology counts, the method and edge orientation histogram, scale-invariant feature transform are similar, and the calculating unlike hog improves accuracy rate based on the density matrix of uniform space.
1. frame-to-frame differences obtains the foreground point information-401 of moving object;
2. extract foreground point hog feature-402, obtain its proper vector;
3. by hog feature extraction to proper vector carry out mating-404 with number of people characteristic model-403, ask for candidate's number of people characteristic area-405 and non-number of people characteristic area-406;
4. judging whether it exists shoulder feature-407 to candidate's number of people characteristic area that the match is successful, when there is shoulder feature, is then number of people characteristic area-408, otherwise is non-number of people characteristic area-409.
3, number of people characteristic area tracking
This method mainly adopts based on coupling and CamShift(self-adaptation mean shift track algorithm continuously) number of people characteristic area tracking technique that combines, for tracking and the location of number of people characteristic area, relate to the step of total method 5..
As Fig. 5, performing step is as follows:
1. the number of people characteristic area-501 that the number of people characteristic area-502 and the present frame that are detected by previous frame detect carries out centre distance and mates-503;
2. when the match is successful, then the target that present frame detects is number of people characteristic area-504, record present frame number of people target area, complete, otherwise enters step 3.;
3. adopt camshift track algorithm to follow the tracks of-505 to previous frame number of people target area, obtain the number of people characteristic area tracing positional of present frame;
4. the present frame number of people characteristic area obtained by camshift as tracking results-506, and is recorded.
Claims (1)
1., based on a people flow rate statistical method for Video Analysis Technology, it is characterized in that comprising the following steps:
1. number of people characteristic model (201) is trained
The gradient orientation histogram feature samples of the number of people is trained, generates number of people characteristic model, namely generate number of people gradient orientation histogram feature samples model;
2. frame-to-frame differences obtains the foreground point (202) of moving object
Frame-to-frame differences method is adopted to obtain the foreground point information of moving object;
3. number of people feature extraction and identification (203)
To the feature extraction algorithm of the image applications Corpus--based Method after frame-to-frame differences, the target in detected image, mates the target detected with sample pattern, tentatively determines the number of people characteristic area in this two field picture;
4. body local feature identification (204)
To the number of people region decision detected, whether it exists shoulder feature, when there is shoulder feature, is then number of people characteristic area, otherwise is non-number of people characteristic area;
5. number of people characteristic area follows the tracks of (205)
Adopt the tracking technique combined based on coupling and target following, the number of people detected is followed the tracks of, obtains the movement locus of pedestrian, and the frame number that number of people characteristic area occurs is added up;
6. flow of the people counting (206)
When number of people target crosses line and region by assigned direction, and when the frame number that number of people target occurs is N, carry out the counting on assigned direction to pedestrian, statistics the party flow of the people upwards, wherein N is natural number, 1<N<200;
Described number of people sample pattern method of setting up performing step is as follows:
A, input number of people feature samples (301), input non-number of people feature samples (302);
B, to input sample carry out hog feature extraction (303), obtain the proper vector of sample;
C, application svm sorter (304), train the proper vector of input;
D, computing by svm, obtain the model (305) of number of people feature;
The detection method performing step of described number of people feature is as follows:
A, frame-to-frame differences obtain the foreground point information (401) of moving object;
B, extraction foreground point hog feature (402), obtain its proper vector;
C, by hog feature extraction to proper vector carry out mating (404) with number of people characteristic model (403), ask for candidate's number of people characteristic area (405) and non-number of people characteristic area (406);
D, candidate's number of people characteristic area that the match is successful is judged whether it exists shoulder feature (407), when there is shoulder feature, be then number of people characteristic area (408), otherwise be non-number of people characteristic area (409);
The performing step of described number of people characteristic area tracking is as follows:
I, the number of people characteristic area (501) that the number of people characteristic area (502) and the present frame that are detected by previous frame detect carries out centre distance and mates (503);
II, when the match is successful, then the target that present frame detects is number of people characteristic area (504), record present frame number of people target area, complete, otherwise enters step 3.;
III, adopt camshift track algorithm to follow the tracks of (505) to previous frame number of people target area, obtain the number of people characteristic area tracing positional of present frame;
IV, the present frame number of people characteristic area obtained by camshift as tracking results (506), and is recorded.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210208666.7A CN102799935B (en) | 2012-06-21 | 2012-06-21 | Human flow counting method based on video analysis technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210208666.7A CN102799935B (en) | 2012-06-21 | 2012-06-21 | Human flow counting method based on video analysis technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102799935A CN102799935A (en) | 2012-11-28 |
| CN102799935B true CN102799935B (en) | 2015-03-04 |
Family
ID=47199034
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210208666.7A Active CN102799935B (en) | 2012-06-21 | 2012-06-21 | Human flow counting method based on video analysis technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102799935B (en) |
Families Citing this family (20)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103021059A (en) * | 2012-12-12 | 2013-04-03 | 天津大学 | Video-monitoring-based public transport passenger flow counting method |
| CN103390172A (en) * | 2013-07-24 | 2013-11-13 | 佳都新太科技股份有限公司 | Estimating method of crowd density under high-density scene |
| CN103473554B (en) * | 2013-08-19 | 2016-12-28 | 上海汇纳信息科技股份有限公司 | Artificial abortion's statistical system and method |
| CN103824114B (en) * | 2014-01-26 | 2016-09-21 | 中山大学 | A kind of pedestrian stream gauge counting method based on cross section traffic statistics and system |
| CN104021605A (en) * | 2014-04-16 | 2014-09-03 | 湖州朗讯信息科技有限公司 | Real-time statistics system and method for public transport passenger flow |
| CN104036250B (en) * | 2014-06-16 | 2017-11-10 | 上海大学 | Video pedestrian detection and tracking |
| CN104063253A (en) * | 2014-07-07 | 2014-09-24 | 无锡智广厦科技有限公司 | Method for automatic statistics of passenger flow, all-in-one machines and distributed system for automatic statistics of passenger flow |
| CN105303191A (en) * | 2014-07-25 | 2016-02-03 | 中兴通讯股份有限公司 | Method and apparatus for counting pedestrians in foresight monitoring scene |
| CN105184258B (en) * | 2015-09-09 | 2019-04-30 | 苏州科达科技股份有限公司 | Method for tracking target and system, human behavior analysis method and system |
| CN105466414B (en) * | 2015-11-17 | 2018-05-29 | 广东欧珀移动通信有限公司 | The recommendation method and server of a kind of navigation routine |
| CN105512640B (en) * | 2015-12-30 | 2019-04-02 | 重庆邮电大学 | A kind of people flow rate statistical method based on video sequence |
| CN106951885A (en) * | 2017-04-08 | 2017-07-14 | 广西师范大学 | A kind of people flow rate statistical method based on video analysis |
| CN108986064B (en) * | 2017-05-31 | 2022-05-06 | 杭州海康威视数字技术股份有限公司 | People flow statistical method, equipment and system |
| CN108364438A (en) * | 2017-12-28 | 2018-08-03 | 武汉烽火众智数字技术有限责任公司 | Method for early warning based on front end perception data and device |
| CN108388883A (en) * | 2018-03-16 | 2018-08-10 | 广西师范大学 | A kind of video demographic method based on HOG+SVM |
| US20230154192A1 (en) * | 2020-10-16 | 2023-05-18 | Boe Technology Group Co., Ltd. | Person flow analysis method and system, electronic device, and readable storage medium |
| CN112818884B (en) * | 2021-02-07 | 2021-11-30 | 中国科学院大学 | Crowd counting method |
| CN113392726B (en) * | 2021-05-26 | 2023-06-02 | 上海飞乐音响股份有限公司 | Method, system, terminal and medium for identifying and detecting head of person in outdoor monitoring scene |
| CN113255567B (en) * | 2021-06-15 | 2021-10-15 | 湖南中惠旅智能科技有限责任公司 | Resource coordination method and system based on intelligent scenic spot |
| CN116245911B (en) * | 2023-02-08 | 2023-11-03 | 珠海安联锐视科技股份有限公司 | Video offline statistics method |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101196991A (en) * | 2007-12-14 | 2008-06-11 | 同济大学 | Close passenger traffic counting and passenger walking velocity automatic detection method and system thereof |
| CN101794382A (en) * | 2010-03-12 | 2010-08-04 | 华中科技大学 | Method for counting passenger flow of buses in real time |
| CN101872431A (en) * | 2010-02-10 | 2010-10-27 | 杭州海康威视软件有限公司 | People flow rate statistical method and system applicable to multi-angle application scenes |
| CN101872422A (en) * | 2010-02-10 | 2010-10-27 | 杭州海康威视软件有限公司 | People flow rate statistical method and system capable of precisely identifying targets |
-
2012
- 2012-06-21 CN CN201210208666.7A patent/CN102799935B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101196991A (en) * | 2007-12-14 | 2008-06-11 | 同济大学 | Close passenger traffic counting and passenger walking velocity automatic detection method and system thereof |
| CN101872431A (en) * | 2010-02-10 | 2010-10-27 | 杭州海康威视软件有限公司 | People flow rate statistical method and system applicable to multi-angle application scenes |
| CN101872422A (en) * | 2010-02-10 | 2010-10-27 | 杭州海康威视软件有限公司 | People flow rate statistical method and system capable of precisely identifying targets |
| CN101794382A (en) * | 2010-03-12 | 2010-08-04 | 华中科技大学 | Method for counting passenger flow of buses in real time |
Non-Patent Citations (2)
| Title |
|---|
| 基于Joint HOG特征复杂场景下的头肩检测;顾炯等;《电子设计工程》;20120430;第20卷(第7期);第58-62页 * |
| 基于梯度方向直方图的行人计数方法研究;刘超;《中国优秀硕士学位论文全文数据库信息科技辑》;20110315(第3期);第7-9、19-32页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102799935A (en) | 2012-11-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102799935B (en) | Human flow counting method based on video analysis technology | |
| CN109644255B (en) | Method and apparatus for annotating a video stream comprising a set of frames | |
| CN107004271B (en) | Display method, display apparatus, electronic device, computer program product, and storage medium | |
| CN105574506A (en) | Intelligent face tracking system and method based on depth learning and large-scale clustering | |
| WO2013189465A2 (en) | Method, device and system for obtaining the number of persons | |
| CN106951885A (en) | A kind of people flow rate statistical method based on video analysis | |
| CN102610102A (en) | Suspect vehicle inspection and control method and system | |
| CN111582122B (en) | System and method for intelligently analyzing behaviors of multi-dimensional pedestrians in surveillance video | |
| CN103824070A (en) | Rapid pedestrian detection method based on computer vision | |
| Yang et al. | Clustering method for counting passengers getting in a bus with single camera | |
| CN102063613A (en) | People counting method and device based on head recognition | |
| Zhang et al. | Crowd density estimation based on statistical analysis of local intra-crowd motions for public area surveillance | |
| Razalli et al. | Emergency vehicle recognition and classification method using HSV color segmentation | |
| CN112232211A (en) | Intelligent video monitoring system based on deep learning | |
| CN104463232A (en) | Density crowd counting method based on HOG characteristic and color histogram characteristic | |
| CN104134067A (en) | Road vehicle monitoring system based on intelligent visual Internet of Things | |
| CN101587592A (en) | Target motion identification system | |
| CN105227918B (en) | A kind of intelligent control method | |
| WO2023071188A1 (en) | Abnormal-behavior detection method and apparatus, and electronic device and storage medium | |
| CN113920585A (en) | Behavior recognition method and device, equipment and storage medium | |
| Kainz et al. | Visual system for student attendance monitoring with non-standard situation detection | |
| Wang et al. | Video image vehicle detection system for signaled traffic intersection | |
| CN104616277B (en) | Pedestrian's localization method and its device in video structural description | |
| Wanjale et al. | Use of haar cascade classifier for face tracking system in real time video | |
| CN115169673A (en) | Intelligent campus epidemic risk monitoring and early warning system and 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 | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |