CN100487739C - Multi-layer real time forewarning system based on the intelligent video monitoring - Google Patents
Multi-layer real time forewarning system based on the intelligent video monitoring Download PDFInfo
- Publication number
- CN100487739C CN100487739C CNB2007101059337A CN200710105933A CN100487739C CN 100487739 C CN100487739 C CN 100487739C CN B2007101059337 A CNB2007101059337 A CN B2007101059337A CN 200710105933 A CN200710105933 A CN 200710105933A CN 100487739 C CN100487739 C CN 100487739C
- Authority
- CN
- China
- Prior art keywords
- video
- module
- early warning
- analysis
- rule
- 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
Images
Landscapes
- Closed-Circuit Television Systems (AREA)
- Image Analysis (AREA)
- Alarm Systems (AREA)
Abstract
The invention relates to an intelligent early warning system, especially a multiple layered real-time early warning system based on target detection, positioning, class recognition of intelligent video analysis and based on network management. The system comprises five portions: a video collecting unit, a video processing unit, an action recognition module based on a rule, a network management platform and an alarming module. The method comprises: 1, a user appointing an early warning rule for every vidicon on the network management platform, realizing a multiple layered real-time early warning mechanism; 2, in the video processing unit, dividing a video sequence into three paths, wherein a path is used for coding and compressing, a path is used for scene analysis, and a path is used for background modeling, the video processing unit has encoding and compressing functions, can obtain a traget track after video analysis; 3, transmitting the traget track via a network to the action recognition module, and performing action recognition according to the rule defined by the user.
Description
Technical field
The present invention relates to the intelligent early-warning system technical field, the multi-level real-time early warning system of particularly a kind of target detection, location, Classification and Identification and management Network Based based on intelligent video analysis.
The information source of real-time early warning system comprises the various sensors and the network information.The early warning of real-time early warning system is to liking the abnormal conditions of the different brackets that takes place under the various scenes.
The sensor that the present invention adopts is a vision sensor.The method that adopts comprises target detection, target following, target classification, scene calibration static scene under, discerns based on the information fusion of reasoning, rule-based behavior.Early warning to as if can define all occasions of early warning rule at many levels.
Background technology
Present early warning system is carried out early warning based on a kind of sensor mostly, and the quantity of information of obtaining is few, have significant limitation.Therefore vision is the main path that the people obtains external information, comprises than sense of hearing sense of touch more information, has bigger application prospect based on the supervisory system of intelligent video analysis than other early warning system.
Existing video monitoring system comprises video acquisition, motion detection and compression storage.Video acquisition has used various camera, comprises CCD camera, CMOS camera, infrared camera etc.Motion detection (also being mobile detection) is to detect the frame that motion is arranged in the video sequence, and not detection and tracking target also is not a kind of intelligent analysis method.The MPEG-4 and the decoding method of standard have H.264 mainly been adopted in compression storage.General video storage is in jumbo hard disk, because the content of storage is too much, video can only keep the limited time.When needs inquiry specific video content, will go to inquire about in the video library time-consuming effort again with artificial method.The more important thing is that when monitoring scene generation abnormal conditions, the early warning in time of traditional supervisory system is normally reported to the police by other modes, the artificial enquiry monitoring video is collected evidence then.Therefore, present supervisory system is mainly used in video recording, inquires about afterwards and collects evidence, and does not have real warning function.
Development along with security protection, in the reality increasing supervisory system will be arranged, with the mode of artificial enquiry incompatibility market to the requirement of supervisory system, this just needs to adopt the method for intellectual analysis, the video of gathering is carried out intellectual analysis, classification and storage, and before abnormal conditions take place, early warning is in time removed a hidden danger.
Early warning in the reality is the branch different brackets, and the advanced warning grade of different occasions is different.Such as, when the people paces up and down in the oil depot periphery, just should early warning and remind the entrance guard to note taking precautions against, when the people crosses enclosure wall near oil depot, should in time report to the police.Just as people's vision is to obtain the main path of external information, can obtain than other early warning system more information based on the supervisory system of visual analysis, by to target detection in the scene and behavioural analysis, can in time detect behaviors such as pacing up and down, cross enclosure wall, early warning as early as possible, the accident that prevents takes place.By the network information management platform, the information of different cameras is shared, and realizes monitoring in a big way.Therefore, the supervisory system based on intelligent video analysis more is applicable to multi-level real-time early warning than other early warning systems.
Summary of the invention
The object of the present invention is to provide a kind of multi-level real-time early warning system based on intelligent video analysis.
Intelligent video analysis can realize target detection tracking, target behavioural analysis and differentiate based on the behavior of User Defined rule.Because the applied environment of video monitoring is ever-changing, this just needs methods of video analyses to have very high robustness, be applicable to different weather, different background and different illumination conditions.Simultaneously, intelligent analysis method must be real-time, so just can reach the purpose of timely early warning, and this just proposes very high requirement to intelligent analysis method.Multi-level early warning system also needs information sharing between the different cameras, realizes the association between the different cameras, and by the rule of user's specified associations, this just require behavioural analysis must be under different occasions accurate and effective all.
Multi-level real-time early warning system based on intelligent video analysis is made up of five parts: video acquisition, video processing unit, rule-based behavior identification, network management platform and alarm module.
Video acquisition module is used to obtain digital video sequences, and compatible different cameras can be as acquisition module as long as can produce the video camera of digital image information.
Video processing unit has adopted background modeling, target to extract, the adaptive parameter control method of illumination, scene are stored from demarcation, target following, track extraction method, video coding and selection.Video processing unit is one of nucleus module of the present invention, and it has the parameter adaptive function to scene, therefore has robustness preferably, target detection when can be used for rainy day, snowy day and night and behavior identification.Simultaneously, adopted convergent-divergent rotation translation invariant feature in the target following, tracking itself also is adaptive, is not subjected to the scene variable effect.
The behavior that rule-based behavior identification module is based under the User Defined rule is differentiated.This module is used for the target trajectory that the analysis video analysis module obtains, according to the rule of upper-layer user's definition, and the grade of early warning, in different target trajectorys, extract legal track, and carry out early warning.
Network management module is used to manage different intelligent analysis terminals, and according to different safe classes multi-level early warning mode is set, and makes all intelligent terminals accept tension management, is convenient to form an area monitoring system.
The signal of alarm module receiving network managing platform comes real-time early warning, also can be the alerting signal from intelligent terminal.Alarm module can be to report to the police by the Network Transmission signal, can be directly to be reported to the police by terminal, also can be to report to the police by different level, determines according to the rule of user's appointment.
In sum, compare with existing video monitoring system, the multi-level real-time early warning system based on intelligent video monitoring that the present invention proposes has following difference:
1) adopted the parameter adaptive control method in the video processing unit, the feature of using in the target following is a convergent-divergent rotation translation invariant feature, and tracking itself also is adaptive, and trajectory analysis has adopted the method based on study and scene calibration.Present existing background modeling, target extract and tracking all is defective, also do not have a kind of method can reach parameter adaptive and are applicable to different weather environments.If make intelligent monitoring practicability, must in method, consider and overcome the influence of weather and environment.The invention solves this problem.
2) adopted behavior recognition methods based on the User Defined rule.The variation of application background has determined regular variation, has practicality widely based on the behavior recognition methods of User Defined rule.
3) adopted the method for multi-level early warning.The function of intelligent terminal often all is related rather than independently in the practical application, and this has level with regard to requiring the warning function between the terminal.Because can there be a spot of false alarm in intelligent monitor system, then can reduce the probability of false alarm based on multi-level early warning system.
Description of drawings
Fig. 1 is the multi-level real-time early warning system based on intelligent video monitoring of the present invention.
Fig. 2 is a video processing unit of the present invention.
Fig. 3 is a rule-based behavior identification module of the present invention.
Embodiment
Based on the multi-level real-time early warning system chart of intelligent video monitoring as shown in Figure 1.Total system is made up of five parts: video acquisition, video processing unit, rule-based behavior identification, multi-level early warning net management platform and alarm module.
Video acquisition module is used to gather video, compatible different video cameras; Video processing unit is used for analyzing and the compression store video; Rule-based behavior identification is used to discern detected target; Multi-level early warning net management platform is used to manage different video cameras, forms the level early warning mechanism; Alarm module is used for transmitting in real time and carrying out warning message.
As shown in Figure 2, video processing unit can be divided into seven subelements again: video encoding module, selection memory module, background modeling module, foreground extracting module, scene analysis module and parameter adaptive adjusting module, target tracking module, target trajectory module.Video encoding module adopts the mpeg standard coding method.Select the result of memory module, only store the video sequence of tracking target, and trace information and audio video synchronization got up and store, help like this inquiring about afterwards and searching for, reduce the demand of supervisory system simultaneously storage space according to intellectual analysis.The background modeling module is used to set up background model, and model parameter is adjusted according to the scene analysis self-adaptation.Foreground extracting module is used for the extraction prospect, and model parameter is adjusted according to the scene analysis self-adaptation.Target tracking module is followed the tracks of the prospect of extracting, and the feature of using in the target following is a convergent-divergent rotation translation invariant feature, and tracking itself also is adaptive.The target trajectory module then comprises the time relevant with target travel, track, size, type, shape, colouring information, and trajectory analysis has adopted the method based on study and scene calibration.
Fig. 3 is based on the behavior identification module of rule, and it comprises Different Rule defined by the user and 2 submodules of behavior identification.Rule is specified by the user through the multi-layer network management platform, is convenient to form an area monitoring like this, also has practicality widely.
Concrete implementing procedure based on the multi-level real-time early warning system of intelligent video monitoring is as follows:
The first, the early warning rule of each video camera of setting also will be carried out scene calibration to video camera in case of necessity on network management platform.Setting by to the early warning rule of each video camera realizes multi-level real-time early warning mechanism.
The second, start-up system and operation.
The 3rd, terminal camera acquisition image sequence is to video processing unit.
The 4th, at video processing unit, video sequence is divided into three the tunnel, the one tunnel and is used for encoding compression, and one the tunnel is used for scene analysis, and one the tunnel is used for background modeling.The function that video processing unit has coding simultaneously and analyzes.Scene analysis extracts the variation characteristic of scene, adopts the method self-adaptation of Adaptive Control Theory to adjust background model and foreground model parameter.Background modeling is set up background model.Foreground extraction is used to extract the foreground moving piece.Used convergent-divergent rotation translation invariant feature in the target following, tracking itself also is adaptive, and trajectory analysis has adopted the method based on study and scene calibration, can reject some false targets or interference like this, obtains real target trajectory.Video processing unit is placed on the terminal, can realize disperseing storage like this and disperse calculating, is easy to network management.
The 5th, obtain after the target trajectory, arrive the behavior identification module through Network Transmission, and carry out behavior identification according to user-defined rule.If terminal itself is exactly one of webserver, target trajectory will directly carry out behavior identification in this locality so.
The 6th, if violate the goal behavior of setting the early warning rule, output alarm signal immediately.
Characteristics of the present invention and effect have:
1) video processing unit has the parameter adaptive ability, and the performance of target following does not change with scene. The greatest problem of video analysis is stability, present existing conventional target detection and tracking side Method all more or less has unsettled defective. The application scenario of intelligent monitoring is ever-changing, has lower Rainy day, snowy day, daytime and evening etc., thus intelligent monitor system must be able to adapt to weather and Environmental evolution just can be practical, and this just requires video analysis that adaptive ability must be arranged.
2) identify based on the behavior of User Defined rule. Rule can be undertaken by network management platform by the user Formulate, concrete application scenario is had more appropriate specific aim, easy to utilize.
3) has multi-level early warning mechanism. Can specify the early warning rule by the user, and a plurality of video cameras are carried out Networking forms a Local Area Network monitoring system, has more widely practicality.
4) reduced requirement to memory space. Only the video sequence that target travel is arranged is stored, can Greatly reduce memory space, be convenient to simultaneously inquire about afterwards, because each video sequence corresponding Individual goal behavior.
5) be convenient to be expanded into the wide area video monitoring system of management Network Based. Along with the monitoring camera number Roll up, coming the management and monitoring video storage with computer is extremely urgent demand, adopts base In the multi-layer network management mode of intelligent video analysis, be the development trend of monitoring system, simultaneously Also be conducive to the retrieval of monitor video.
6) realization of system can be one of following three kinds of modes: hardware is realized, software is realized or software and hardware In conjunction with realizing.
Claims (8)
1, a kind of multi-level early warning system based on intelligent video monitoring, it is characterized in that, form by five modules: video acquisition module, video processing unit, rule-based behavior identification module, network management platform and alarm module, its module annexation is: video acquisition module, video processing unit, rule-based behavior identification module, network management platform and alarm module physically are connected in series successively, and the function of five modules is respectively:
Video acquisition module is used to obtain digital video sequences, and compatible different cameras can be as acquisition module as long as can produce the video camera of digital image information;
Video processing unit comprises: video encoding module, selection memory module, background modeling module, foreground extracting module, scene analysis module and parameter adaptive adjusting module, target tracking module, target trajectory module;
Video encoding module adopts the mpeg standard coding method, selects the result of memory module according to intellectual analysis, only stores the video sequence of tracking target, and trace information and audio video synchronization got up and stores; The background modeling module is used to set up background model, and model parameter is adjusted according to the scene analysis self-adaptation; Foreground extracting module is used for the extraction prospect, and model parameter is adjusted according to the scene analysis self-adaptation; Target tracking module is followed the tracks of the prospect of extracting; The target trajectory module then comprises the time relevant with target travel, track, size, type, shape, colouring information, and trajectory analysis has adopted the method based on study and scene calibration;
Rule-based behavior identification module is used for behavior identification, according to the rule of user upper strata definition, and the grade of early warning, in different target trajectorys, extract legal track, and carry out early warning;
Network management platform is used to manage different intelligent analysis terminals, and according to different safe classes multi-level early warning mode is set, and makes all intelligent terminals accept tension management, is convenient to form an area monitoring system;
Alarm module, the signal of receiving network managing platform comes real-time early warning;
The allomeric function of system is: manage all the other four modules by network management platform, form multi-level intelligent video monitoring system, each video processing unit detects the target trajectory in zone to be detected adaptively, identify abnormal behaviour by identification module then, report to the police by alarm module again based on behavior; Its execution sequence in logic is to set the early warning rule of each video camera earlier on network management platform, then by video acquisition module collection video, video analysis and storage, rule-based behavior identification, last output alarm signal.
2, according to the multi-level early warning system based on intelligent video monitoring of claim 1, it is characterized in that video processing unit wherein, the background modeling and the prospect parameter extraction that are used for video analysis are self-adaptation adjustment, the feature that adopts in the tracking is that the convergent-divergent rotation is translation invariant, tracking itself is adaptive, trajectory analysis is based on study and scene calibration, and video storage is selected to store according to the video analysis result.
3, according to the multi-level early warning system based on intelligent video monitoring of claim 1, its rule-based behavior identification module is based on the User Defined rule.
4, according to the multi-level early warning system based on intelligent video monitoring of claim 1, its network management platform is multi-level real-time, can form the platform of an area monitoring.
5, according to the multi-level early warning system based on intelligent video monitoring of claim 1, its system expands to the wide area network monitor supervision platform.
6, according to the multi-level early warning system based on intelligent video monitoring of claim 1, its system applies is in the video frequency searching based on behavior, target or incident.
7, according to the multi-level early warning system based on intelligent video monitoring of claim 1, its system applies is in any occasion that video camera is housed.
8, according to the multi-level early warning system based on intelligent video monitoring of claim 1, the realization of its system is one of following three kinds of modes: hardware is realized, software is realized or software and hardware combining realizes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007101059337A CN100487739C (en) | 2007-06-01 | 2007-06-01 | Multi-layer real time forewarning system based on the intelligent video monitoring |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CNB2007101059337A CN100487739C (en) | 2007-06-01 | 2007-06-01 | Multi-layer real time forewarning system based on the intelligent video monitoring |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101055662A CN101055662A (en) | 2007-10-17 |
| CN100487739C true CN100487739C (en) | 2009-05-13 |
Family
ID=38795483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNB2007101059337A Expired - Fee Related CN100487739C (en) | 2007-06-01 | 2007-06-01 | Multi-layer real time forewarning system based on the intelligent video monitoring |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN100487739C (en) |
Families Citing this family (24)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101320505B (en) * | 2008-07-04 | 2010-09-22 | 浙江大学 | Partition video monitoring method based on multipath network video stream parallel processing |
| CN101448145A (en) * | 2008-12-26 | 2009-06-03 | 北京中星微电子有限公司 | IP camera, video monitor system and signal processing method of IP camera |
| US20100208063A1 (en) * | 2009-02-19 | 2010-08-19 | Panasonic Corporation | System and methods for improving accuracy and robustness of abnormal behavior detection |
| CN101674466B (en) * | 2009-10-15 | 2011-01-05 | 上海交通大学 | Multi-information fusion intelligent video monitoring fore-end system |
| CN102860003A (en) * | 2010-03-09 | 2013-01-02 | 常州环视高科电子科技有限公司 | A surveillance system and method |
| CN101867793A (en) * | 2010-05-14 | 2010-10-20 | 蔡晓东 | Distribution type intelligent video searching system and using method |
| CN102479419A (en) * | 2010-11-29 | 2012-05-30 | 上海真新资产经营管理有限公司 | Automatic early warning development system and method thereof |
| CN102005106B (en) * | 2010-12-01 | 2012-08-15 | 大连理工大学 | Shipping on-duty monitoring instrument |
| CN102752574B (en) * | 2011-04-18 | 2015-01-28 | 中兴通讯股份有限公司 | Video monitoring system and method |
| CN102811343B (en) * | 2011-06-03 | 2015-04-29 | 南京理工大学 | Intelligent video monitoring system based on behavior recognition |
| EP2772049A4 (en) * | 2011-10-24 | 2015-06-17 | Intel Corp | Multiple stream processing for video analytics and encoding |
| CN102509367A (en) * | 2011-10-28 | 2012-06-20 | 江苏省电力公司常州供电公司 | Three-dimensional (3D) simulation polling system of electric power automation motor room |
| CN103136763B (en) * | 2011-12-02 | 2017-07-21 | 富泰华工业（深圳）有限公司 | Electronic installation and its method for detecting the abnormal paragraph of video sequence |
| CN102752582A (en) * | 2012-06-19 | 2012-10-24 | 杭州海康威视系统技术有限公司 | Intelligent video analysis system and method |
| CN104299355A (en) * | 2014-09-28 | 2015-01-21 | 厦门蓝斯通信股份有限公司 | Vehicle monitoring method and system based on video intelligent recognition |
| CN104301686A (en) * | 2014-10-27 | 2015-01-21 | 青岛宝微视控信息技术有限公司 | Intelligent video analyzing system and method |
| CN104363426A (en) * | 2014-11-25 | 2015-02-18 | 深圳北航新兴产业技术研究院 | Traffic video monitoring system and method with target associated in multiple cameras |
| CN106230833B (en) * | 2016-08-04 | 2019-04-16 | 李东 | The method of network service is provided |
| CN107274618A (en) * | 2017-07-29 | 2017-10-20 | 安徽博威康信息技术有限公司 | A kind of monitor and alarm system based on fixation and recognition |
| CN107403535A (en) * | 2017-07-29 | 2017-11-28 | 安徽博威康信息技术有限公司 | A kind of monitor and alarm system of identity-based identification |
| CN110503812A (en) * | 2019-08-23 | 2019-11-26 | 陈�峰 | A kind of flammability hazard source behavior monitoring system and method based on convolutional neural networks |
| CN111338669B (en) * | 2020-02-17 | 2023-10-24 | 深圳英飞拓仁用信息有限公司 | Method and device for updating intelligent function in intelligent analysis box |
| CN113255534B (en) * | 2021-05-28 | 2022-08-12 | 河北幸福消费金融股份有限公司 | Early warning method, system, device and storage medium based on video image analysis |
| CN113570580B (en) * | 2021-07-30 | 2024-06-07 | 重庆予胜远升网络科技有限公司 | Power equipment component loosening identification system and method based on machine vision |
-
2007
- 2007-06-01 CN CNB2007101059337A patent/CN100487739C/en not_active Expired - Fee Related
Non-Patent Citations (4)
| Title |
|---|
| 基于NMI特征的目标识别与跟踪. 门蓬涛，张秀彬，张峰，孙志旻，吴炯.微计算机信息，第20卷第3期. 2007 |
| 基于NMI特征的目标识别与跟踪. 门蓬涛，张秀彬，张峰，孙志旻，吴炯.微计算机信息，第20卷第3期. 2007 * |
| 智能化视频分析技术探讨. 耿征.中国安防. 2007 |
| 智能化视频分析技术探讨. 耿征.中国安防. 2007 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101055662A (en) | 2007-10-17 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN100487739C (en) | Multi-layer real time forewarning system based on the intelligent video monitoring | |
| CN110428522B (en) | Intelligent security system of wisdom new town | |
| CN102811343B (en) | Intelligent video monitoring system based on behavior recognition | |
| Albiol et al. | Detection of parked vehicles using spatiotemporal maps | |
| CN100555348C (en) | Intelligent video monitoring system of bank self-aid apparatus | |
| CN103108159B (en) | Electric power intelligent video analyzing and monitoring system and method | |
| CN101493979B (en) | Method and instrument for detecting and analyzing intelligent network vision target | |
| CN101123721A (en) | An intelligent video monitoring system and its monitoring method | |
| CN101854516A (en) | Video monitoring system, video monitoring server and video monitoring method | |
| CN106408833A (en) | Perimeter intrusion detection method and system | |
| CN105426820A (en) | Multi-person abnormal behavior detection method based on security monitoring video data | |
| CN201114536Y (en) | A video monitoring system | |
| CN103069434A (en) | Multi-mode video event indexing | |
| CN111144291A (en) | Method and device for distinguishing personnel invasion in video monitoring area based on target detection | |
| CN102170563A (en) | Intelligent person capture system and person monitoring management method | |
| CN102306304A (en) | Face occluder identification method and device | |
| CN207164939U (en) | Periphery intrusion preventing device based on target position information fusion | |
| CN103632479A (en) | Home-based care for the aged security and protection system based on human motion trajectory analysis | |
| CN102291574A (en) | Complicated scene target movement tracking system based on embedded technique and light transmission and monitoring method thereof | |
| CN112163572A (en) | Method and device for identifying object | |
| CN103281518A (en) | Multifunctional networking all-weather intelligent video monitoring system | |
| CN109948474A (en) | AI thermal imaging all-weather intelligent monitoring method | |
| CN112070212A (en) | Artificial intelligence CNN, LSTM neural network dynamic identification system | |
| CN102651813A (en) | Intelligent video monitoring and advertising system | |
| CN108074363A (en) | Perimeter protection monitoring identifying system based on depth camera |
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 | ||
| C41 | Transfer of patent application or patent right or utility model | ||
| TR01 | Transfer of patent right |
Effective date of registration: 20091211 Address after: Room D, building 205H, No. 2, information road, Beijing, Haidian District, China: 100085 Co-patentee after: Beijing neutralization Signal Technology Co., Ltd. Patentee after: Beijing Hui Da Tong Ye Technology Co., Ltd. Co-patentee after: Zou Huihua Address before: Room D, building 205H, No. 2, information road, Beijing, Haidian District, China: 100085 Co-patentee before: Hong Yiping Patentee before: Beijing Hui Da Tong Ye Technology Co., Ltd. Co-patentee before: Zou Huihua |
|
| DD01 | Delivery of document by public notice |
Addressee: Beijing Huida Tongye Technology Co., Ltd. Document name: Notification to Pay the Fees |
|
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20090513 Termination date: 20160601 |