Summary of the invention
The purpose of the present invention is intended at least solve one of described technological deficiency.
To this end, it is an object of the invention to propose the video analysis detection method of a kind of crowd massing and movement and system,Crowd massing in environment and crowd are run and detects, thus send alarm signal in time when finding above-mentioned situation, logicalKnow that monitoring personnel gather measure and process.
To achieve these goals, the embodiment of one aspect of the present invention provides a kind of crowd massing and the video analysis of movementDetection method, comprises the steps:
Step S1, by the video figure of each position in the video front-end equipment acquisition monitoring region being arranged on monitoring regionPicture；
Step S2, is stored described video image to database server by switch transmission, and sends to answeringIntelligent video analysis is carried out with server；
Step S3, is analyzed face characteristic by described application server from described video image, and according to default colonyAggregation algorithms, statistician's quantity, and judge whether the personnel amount in described monitoring region exceedes threshold value；
Step S4, when judging that monitoring region personnel amount exceedes threshold value, by described application server to monitoring personnel'sTerminal unit alert, prompting monitoring personnel carry out concern process, and by the video image in this region in real time in displayShow on screen.
Further, in described step S3, farther include: after analyzing face characteristic from described video image,Determine whether the action behavior of personnel, when the personnel of detecting run and run the quantity of personnel exceed threshold value time, sendWarning message, prompting monitoring personnel carry out concern process.
Further, in described step S4, the terminal unit of described monitoring personnel obtains regarding of described monitoring region in real timeFrequently image, and receive described warning message, to point out user to take treatment measures in time.
Further, in described step S4, described warning message is sound and light alarm prompting.
The embodiment of the present invention also proposes the video analysis detecting system of a kind of crowd massing and movement, including: video frontEquipment, database server, switch, application server and display device, wherein,
Described video front-end equipment includes: multiple video cameras, and the plurality of video camera is distributed in multiple positions in monitoring regionPut place, for gathering the video image of each position；
Described switch is connected with described video front-end equipment, described database server and described application server, usesIn described video image is sent to described database server and described application server；
Described database server is used for storing described video image and backing up；
Described application server is for analyzing face characteristic from described video image, and assembles calculation according to default colonyMethod, statistician's quantity, and judge whether the personnel amount in described monitoring region exceedes threshold value, when judging monitoring area peopleWhen quantity exceedes threshold value, by described application server to the terminal unit alert of monitoring personnel, point out monitoring personnelCarry out concern process；
Described display device is connected with described application service district, for by the video image in this region the most on a display screenShow.
Further, after described application server analyzes face characteristic from described video image, people is determined whetherThe action behavior of member, when the personnel of detecting run and run the quantity of personnel exceed threshold value time, alert, promptingMonitoring personnel carry out concern process.
Further, when described monitoring region is positioned at outdoor, the installation site distance detection region of each described video camera10 meters～15 meters of 20 meters～40 meters, angle 45 degree or 45-60 degree, height distance ground；
When described monitoring region is positioned at indoor, 8 meters of the installation site distance detection region of each described video camera～20Rice, angle 45 degree or 45-60 degree, 5 meters～10 meters of height distance ground.
Further, the terminal unit of described monitoring personnel obtains the video image in described monitoring region in real time, and receives instituteState warning message, to point out user to take treatment measures in time.
Further, the terminal unit of described monitoring personnel is mobile phone or panel computer.
Further, described warning message is sound and light alarm prompting.
Crowd massing according to embodiments of the present invention and the video analysis detection method of movement and system, can be according to regardingFrequency gather data analysis, the crowd massing in environment and crowd are run and detect, thus find above-mentioned situation time andTime send alarm signal, notice monitoring personnel gather measure and process.
A set of intelligence that the present invention continues to optimize design based on multiple Intelligent Recognition algorithm knows serial application product, for userProvide Intelligent Recognition, intelligence evidence obtaining, early warning in advance, dispose afterwards, three grades of linkages, craft warning, mobile terminal application management, in advanceThe multiple application services such as alarming information classification propelling movement, it is possible to by the intellectual analysis to video data before event occurs, in advanceSend early warning, and realize three grades of linkage early warning according to the degree of danger of event, the image of event can be distributed to personnel at different levels and moveIn dynamic terminal, win the quality time for event handling, the impact of public or private property and people's life's safety is reduced by eventTo minimum level.
Aspect and advantage that the present invention adds will part be given in the following description, and part will become from the following descriptionObtain substantially, or recognized by the practice of the present invention.
Detailed description of the invention
Embodiments of the invention are described below in detail, and the example of described embodiment is shown in the drawings, the most from start to finishSame or similar label represents same or similar element or has the element of same or like function.Below with reference to attachedThe embodiment that figure describes is exemplary, it is intended to is used for explaining the present invention, and is not considered as limiting the invention.
As it is shown in figure 1, the video analysis detection method of the crowd massing of the embodiment of the present invention and movement, including walking as followsRapid:
Step S1, by the video figure of each position in the video front-end equipment acquisition monitoring region being arranged on monitoring regionPicture；
Specifically, the data acquisition program in video front-end equipment gather data (xml, picture and video segment), logicalAfter crossing process, data are stored respectively in data base and storage server.Can also realize accessing third party by interface and be safelySystem or rays safety detection apparatus detection information.
Step S2, is stored video image to database server by switch transmission, and sends to applying clothesBusiness device carries out intelligent video analysis.
With reference to Fig. 4, analysis result is sent to data acquisition program, data acquisition by video analytics server with xml dataProgram gathers picture and video segment by xml from video analytics server.Data after data acquisition program will process are dividedCun Chu be with data base and storage server.
Step S3, is analyzed face characteristic by application server from video image, and according to default colony aggregation algorithms,Statistician's quantity, and judge whether the personnel amount monitored in region exceedes threshold value.
Step S4, when judging that monitoring region personnel amount exceedes threshold value, by application server to the terminal of monitoring personnelEquipment alert, prompting monitoring personnel carry out concern process, and by the video image in this region the most on a display screenShow.
The terminal unit of monitoring personnel obtains the video image in monitoring region in real time, and receives warning message, to point out useTreatment measures are taked at family in time.Wherein, warning message is sound and light alarm prompting.
Below the process of crowd massing detection is illustrated；Crowd massing algorithm is all in the range of realizing detection video cameraThe colony of people assembles, and in actual applications, colony's gathering can be judged by algorithm routine according to certain data threshold values, works as calculationMethod thinks that when reaching certain threshold values, system will be reported to the police.When the colony of detection crowd assembles, such as the gathering of crowd, when detection regardsFrequently having exceeded the behavior setting threshold values in scene, system is reported to the police automatically.
Range of application: be applicable to the places such as Grand Place relatively wide, the visual field that photographic head is clapped also compares the widest scene, such as park,Square, important office of enterprise doorway, building (road, highway is absent from the scene in the range of scape).For different photographic head position andAngle, parameter used also has bigger difference.
Outdoor environment requires: monitoring region distance photographic head 20m～40m；Daytime, light stabilisation；At night, there is bright sourceAnd stable, light-source brightness can not suddenly change；Angle 45 degree or 45-60 degree；Height distance ground 10m～15m.
Indoor environment requires: monitoring region distance photographic head 8m～20m；Light stabilisation or have bright source and stable, lightSource brightness can not suddenly change；Angle 45 degree or 45-60 degree；Height distance ground 5m～10m.
Additionally, after analyzing face characteristic from video image, determine whether the action behavior of personnel, when detectingPersonnel's run and run quantity of personnel is when exceeding threshold value, and alert, prompting monitoring personnel carry out concern process.
Specifically, crowd run algorithm realize detection video camera in the range of proprietary colony run, in actual applications,Group behavior can be judged by algorithm routine according to certain data threshold values, when algorithm think reach certain threshold values time, systemWill report to the police.The crowd of detection crowd is run, and as a direction oriented in crowd or multiple directions are run, system is reported automaticallyAlert.
Range of application: be applicable to the places such as Grand Place relatively wide, the visual field that photographic head is clapped also compares the widest scene, such as park,Square, important office of enterprise doorway, building (road, highway is absent from the scene in the range of scape).For different photographic head position andAngle, parameter used also has bigger difference.
Outdoor environment requires: monitoring region distance photographic head 20m～40m；Daytime, light stabilisation；At night, there is bright sourceAnd stable, light-source brightness can not suddenly change；Angle 45 degree or 45-60 degree；Height distance ground 10m～15m.
Indoor environment requires: monitoring region distance photographic head 8m～20m；Light stabilisation or have bright source and stable, lightSource brightness can not suddenly change；Angle 45 degree or 45-60 degree；Height distance ground 5m～10m.
As in figure 2 it is shown, the video analysis detecting system of the crowd massing of the embodiment of the present invention and movement, including: before videoEnd equipment 1, database server 2, switch 3, application server 4 and display device 5.
Concrete, video front-end equipment 1 includes: multiple video cameras, and multiple video cameras are distributed in multiple positions in monitoring regionPut place, for gathering the video image of each position.
When monitoring region and being positioned at outdoor, 20 meters～40 meters of the installation site distance detection region of each video camera, angle45 degree or 45-60 degree, 10 meters～15 meters of height distance ground；
When monitoring region and being positioned at indoor, 8 meters～20 meters of the installation site distance detection region of each video camera, angle 45Degree or 45-60 degree, 5 meters～10 meters of height distance ground.
Switch 3 is connected with video front-end equipment 1, database server 2 and application server 4, for by video imageSend to database server 2 and application server 4.Database server 2 is for store video images and backs up.
Specifically, the data acquisition program in video front-end equipment 1 gather data (xml, picture and video segment), logicalAfter crossing process, data are stored respectively in data base and storage server.Can also realize accessing third party by interface and be safelySystem or rays safety detection apparatus detection information.
First, analysis result is sent to data acquisition program, data acquisition program by video analytics server with xml dataFrom video analytics server, picture and video segment is gathered by xml.Data after data acquisition program will process are deposited respectivelyStorage and data base and storage server.
Application server 4 is for analyzing face characteristic from video image, and according to default colony aggregation algorithms, adds upPersonnel amount, and judge whether the personnel amount monitored in region exceedes threshold value, when judging that monitoring region personnel amount exceedes thresholdDuring value, application server 4 carry out concern process to the terminal unit alert of monitoring personnel, prompting monitoring personnel.
In one embodiment of the invention, application server 4 includes: Distributor, reception server, algorithm serviceDevice.
Wherein, after application server 4 analyzes face characteristic from video image, determine whether the action row of personnelFor, when the personnel of detecting run and run the quantity of personnel exceed threshold value time, alert, prompting monitoring personnel enterRow concern processes.
The terminal unit of monitoring personnel obtains the video image in monitoring region in real time, and receives warning message, to point out useTreatment measures are taked at family in time.In one embodiment of the invention, the terminal unit of monitoring personnel is mobile phone or panel computer.
Front end monitoring collecting device transmits the video to backstage by network and is analyzed, and background work personnel can pass through,PC, mobile terminal (mobile phone, crime scene investigation device), giant-screen check relevant video and early warning information.Wherein, warning message is sound and light alarmPrompting.
Display device 5 is connected with application service district, for being shown the most on a display screen by the video image in this regionShow.
As it is shown on figure 3, seven levels of system architecture, it is operating system layer, database layer, web container layer respectively, runs ringBorder layer, technology layer, Internet, access browse layer.
Below the software support platform of the present invention is illustrated:
1, Oracle11g data base: provide reliable, safe data management function with the crucial business event of reply andLine issued transaction is applied.Oracle advanced security option provide transparent data encryption, can to lane database storage data andThe data of network transmission are encrypted.
2, operating system (windows/linux): as operating system on authorization server is relevant soft for runningPart.
3, resource service (WebServer): provide resource information for other service by (resource service) operation system,It is operation system and the bridge of other service.
4, distribution service (VideoServer): the Video stream information of hardware device collection is forwarded to Analysis Service etc..
5, website service (tomcat): the use of the Internet end management system is provided as web container.
6, middleware (ESB): for integrated third party software and the interface of hardware.
7, electronic chart (GIS): user provides integral platform electronic map function.
8, storage service (StorageServer): realize the storage function of history video record.
9, program (FKCounter) is received: realize the timely reception of early warning event information, and operation system real-time exhibition.
10, picture publisher server (nginx): for issuing the picture of early warning event.
11, video playback plug-in unit (OCX): realizing operation system can be with displaying live view video playback.
12, video monitoring client terminal (VideoClient): one control is provided to client, accesses the visual of headend equipmentChange platform.By client, user can control equipment The Cloud Terrace, 3D location, viewing with the video information of displaying live view all devicesHistory video record, browses video scheme broadcasting etc..
13, operation system: provide to client consult in time, process from front end equipment Inspection to Realtime Alerts event flatPlatform.
14, defrag .rar: make regular check on and clear up video record and the picture file of storage, prevent disk space notFoot.
The present invention can dispose on authorization server, and supports the soft or hard system integration customized, and compares other and is embedded in shootingFor the product that machine or soft or hard combine, extensibility is strong, can be according to some function of the increase in demand of user；The algorithm usedAccording to substantial amounts of contrast, test, in the case of environment reaches the requirement of products application, the accuracy rate of algorithm embeds higher than otherThe product that video camera or soft or hard combine；Algorithm, client etc. has broken string to automatically reply function；Whole system can be compatibleThe systems such as windows2003/xp/win7/win8.
Crowd massing according to embodiments of the present invention and the video analysis detection method of movement and system, can be according to regardingFrequency gather data analysis, the crowd massing in environment and crowd are run and detect, thus find above-mentioned situation time andTime send alarm signal, notice monitoring personnel gather measure and process.
A set of intelligence that the present invention continues to optimize design based on multiple Intelligent Recognition algorithm knows serial application product, for userProvide Intelligent Recognition, intelligence evidence obtaining, early warning in advance, dispose afterwards, three grades of linkages, craft warning, mobile terminal application management, in advanceThe multiple application services such as alarming information classification propelling movement, it is possible to by the intellectual analysis to video data before event occurs, in advanceSend early warning, and realize three grades of linkage early warning according to the degree of danger of event, the image of event can be distributed to personnel at different levels and moveIn dynamic terminal, win the quality time for event handling, the impact of public or private property and people's life's safety is reduced by eventTo minimum level.
In order to adapt to user's request, the present invention has environment learning ability, can carry according to constantly improving of security protection identification storehouseHigh system recognition rate, improves the security protection work efficiency of relevant departments, preferably ensures life and the property safety of broad masses, dimensionProtect social stability.
The crowd massing of the embodiment of the present invention and the video analysis detection method of movement and system, have following useful effectReally:
1, comprehensive
Most current Intelligent Recognition manufacturer only achieves the module of or a few Intelligent Recognition, it is impossible to meet userThe demand of growing lifting safe class.Just the product of multiple commercial vendors all can only be filled one when user has multiple demandTime, different functions realizes on different platforms.And for article, colony personnel, environment, all the 11 of photographic head itselfCategorles analytic function all realizes on a platform, can meet the demand of most of client, to easy to use from the point of view of user.
2, data evidence obtaining is quick
Most current intelligence manufacturer is by preposition for the algorithm of Intelligent Recognition, and this results in evidence obtaining after an event occurs and is strandedDifficulty, and Intelligent Recognition platform is owing to being to dispose on backstage, it is possible to analyze the video of storage easily, location needs take rapidlyThe data of card, save substantial amounts of time cost for client.And support multistage linking, data are distributed.
3, three grades of linkage early warning
System can realize patrol personnel's handheld terminal, monitoring center system platform, three grades of linkages of headquarter's system platform,Personnel for user's many levels provide decision support, it is achieved early warning is disposed in advance, event is compressed among cradle.
4, sustainable development
The constantly identification storehouse of sophisticated systems, improves discrimination, simultaneously, it is possible to know according to customer demand customized development newly intelligenceOther algorithm meets different demands, keeps permanent sustainable development in following application process.
5, bidirectional alarming
The present invention can realize the bidirectional alarming mechanism of video analysis early warning and on-the-spot patrol personnel's manual alarm, in shootingHead blind area, alarm image beamed back early warning platform by on-the-spot patrol personnel manually form, and platform commanding is according to image feelingsCondition arranges disposing task.Thus form, by photographic head and Field Force, the early warning commander covered without dead angle and dispose network.
6, protection original investment
Owing to Intelligent Recognition platform is a set of pure software system, it is arranged in the backstage of video monitoring system, so without rightThe video monitoring system that user has built up carries out transforming or removing reconstruction, it is possible to be effectively protected the original investment of client, phaseTo current domestic most monitoring manufacturers by preposition for the algorithm pattern to photographic head, Intelligent Recognition platform can maximize protectionThe original investment of client.
7, economy
The Intelligent Recognition platform of research and development can be adjusted flexibly the region of Intelligent Recognition, and this region after client deploymentThe content identified.The such as region of some camera for supervising the square shooting, what daytime identified is probably the behavior knowledge of colony or individuality, the most likely do not identify evening is the identification of doubtful dangerous materials or legacy.Traditional monitoring manufacturer can only rely on increase to take the photographCamera, increases equipment investment and realizes, Intelligent Recognition platform then only need to arrange in software it is achieved that save greatlyThe about engineering of user puts into, equipment investment, and human input etc. has a high economy.Urban safety aspect for another example, noThe same time period needs the region of Intelligent Recognition to be different, and traditional solution installs substantial amounts of photographic head, Lao Min exactly additionalLose money.Intelligent Recognition platform has only to arrange in software and can realize, and possesses high economy.
Simultaneously as Intelligent Recognition platform is, by computer, video is carried out intellectual analysis, it is possible to reduce substantial amounts of on dutyPersonnel, for using unit to save substantial amounts of human cost.Possesses the highest economy.
8, practicality
Intelligent Recognition platform has extremely perfect early warning mechanism, the most can realize comprising article identification, personnelThe 11 big functions such as Activity recognition, Context awareness, photographic head oneself state identification, can be according to analytical data by contingent thingPart is reported to the police, and allows user carry out the disposal being correlated with when event bud, with the loss avoiding event to cause after occurring.
9, mobile interchange application
Use for the convenience of the user, in addition to the platform being arranged on Surveillance center, the present invention have also been devised cell phone customerEnd (crime scene investigation device) application, sets according to scene patrol disposal personnel, floor manager personnel, the different rights of personnel of action commander in chiefPut different inquiry, push, seek help, the function such as transfer, user can be applied by mobile phone terminal and realize efficiently and commander peopleMember, the two-way interaction of command centre, greatly enhance scope and the convenience of system application.
10, wechat pushes application
The support or the reinforcing personnel that do not install the mobile phone (crime scene investigation device) of native system for convenience pick the early warning that system pushesInformation, system specialized designs wechat pushing module, early warning information can be pushed to the micro-signal specified.And can be by paying close attention to wechatPublic number, it is achieved public number pushes automatically.It is multiple interior that propelling movement content contains early warning event property, early warning event urgency level etc.Hold.User can arrange propelling movement content on platform, pushes the multiple personalizations such as the frequency and arranges.
11, high efficiency
A set of platform of intellectual analysis based on video image, platform is realizing the basis of the intellectual analysis to video imageOn by different scenes, varying environment arranges different parameters and realizes the early warning in advance to event.Early warning information passes through platformThree grades of linkage early warning mechanisms be automatically distributed to the client of different persons liable, platform can also pass through correlation function simultaneouslyModule realizes emergency command.The quick disposal to event is realized by the linkage of each functional module of whole platform.
Using pure software form to realize, algorithm uses OpenCV to combine the mode of c/c++ and realizes, OpenCV be one based onThe cross-platform computer vision library that (increasing income) issues, may operate in Linux, Windows and MacOS operating system.It is lightMagnitude and efficiently, be made up of a series of C function and a small amount of C++ class, provide the language such as Python, Ruby, MATLAB simultaneouslyInterface, it is achieved that a lot of general-purpose algorithms in terms of image procossing and computer vision.The vision processing algorithm that OpenCV providesThe abundantest, and its part writes with C language, it is not necessary to add and new outside support that compiling link that can also be complete generatesExecution program.Portability of the present invention is high, and stability is strong, adapts to various possible environment, and technology autgmentability is strong.
Whole system after deployment, if there is early warning event, sends warning message automatically, and user can check and locate in timeReason.System availability is higher, is suitable for various environment, and a video camera can load several algorithm simultaneously, automatically generate warningInformation, intuitive is strong, saves substantial amounts of manpower and materials, safeguards simple, and expansible new function.
In the description of this specification, reference term " embodiment ", " some embodiments ", " example ", " specifically showExample " or the description of " some examples " etc. means to combine this embodiment or example describes specific features, structure, material or spyPoint is contained at least one embodiment or the example of the present invention.In this manual, to the schematic representation of above-mentioned term notNecessarily refer to identical embodiment or example.And, the specific features of description, structure, material or feature can be anyOne or more embodiments or example in combine in an appropriate manner.
Although above it has been shown and described that embodiments of the invention, it is to be understood that above-described embodiment is exampleProperty, it is impossible to be interpreted as limitation of the present invention, those of ordinary skill in the art is without departing from the principle of the present invention and objectiveIn the case of above-described embodiment can be changed within the scope of the invention, revise, replace and modification.The scope of the present inventionLimited by claims and equivalent thereof.