CN105979210A - Pedestrian identification system based on multi-ball multi-gun camera array - Google Patents
Pedestrian identification system based on multi-ball multi-gun camera array Download PDFInfo
- Publication number
- CN105979210A CN105979210A CN201610396862.XA CN201610396862A CN105979210A CN 105979210 A CN105979210 A CN 105979210A CN 201610396862 A CN201610396862 A CN 201610396862A CN 105979210 A CN105979210 A CN 105979210A
- Authority
- CN
- China
- Prior art keywords
- gunlock
- target
- server
- ball machine
- ball
- 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.)
- Granted
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
Abstract
The invention discloses a pedestrian identification system based on a multi-ball multi-gun camera array, comprising at least one gun-shaped machine, a plurality of ball machines affiliated to the gun-shaped machine and a server; the gun-shaped machine is used for collecting video images in a whole monitoring scene; the server is used for detecting and tracking all motion objects in the video image and controlling work of all ball machines corresponding to the gun-shaped machine when an abnormal motion target is detected; the ball machine is used for regulating collection parameters of its own and collecting a resolution video image corresponding to an abnormal motion object under control of the server; and the server is also used for extracting multi-dimension characteristics in the high resolution video images of all abnormal motion objects collected by the ball machine and storing the multi-dimension characteristics in a corresponding view database for follow-up comparison and identification of the abnormal targets. The multi-ball multi-gun camera array pedestrian identification system solves problems of wide range monitoring and local high definition monitoring, can capture images of multiple visual angles through the ball machine, guarantees integrity and effectiveness of information and improves follow-up identification and comparison performance.
Description
Technical field
The present invention relates to technical field of video monitoring, take the photograph based on the many balls of multiple gun in particular to one
The pedestrian of camera array identifies system.
Background technology
Along with the development of science and technology, the security and guard technology of modernization is the most increasingly towards networking, numeral
Change, intelligent, modular development, the most various advanced security protection means mutually embed, link, complementary,
Modern security product is reached its maturity practicality.It is usually used in residential block such as video monitoring system and road conditions are handed over
The scenes such as logical monitoring, are carried out so that the living safety and traffic order that ensure resident is ruly.
At present, the video monitoring equipment being preinstalled with video monitoring system is applied to each corner of society,
Bringing to social security and develop by leaps and bounds, it act as " eyes " of monitoring personnel, by above-mentioned video
Monitoring device, it is possible to the video image that convenient investigation when and where to be investigated is corresponding, for thing afterwards
Key effect is played in part investigation, also cracks the reliable evidence of the science of providing for case.
For being arranged on the video monitoring equipment of residential block, one high definition gunlock of general employing and server,
Covered the video image of whole residential block by high definition gunlock collection, and gathered by server analysis gunlock
Face characteristic in video image, and the analysis result of this video image and corresponding face characteristic is entered
Row preserves, in order to be subsequently used for the investigation of event afterwards and as deposit evidence.
But, the resolution of the video image gathered by high definition gunlock of the prior art is relatively low, when
When it is used for identifying comparison target as evidence, accuracy of identification is poor.
Summary of the invention
In view of this, the purpose of the embodiment of the present invention is to provide a kind of based on multiple gun many balls video camera battle array
The pedestrian of row identifies system, it is possible to solves large-range monitoring and the problem of local high-definition monitoring, improves
Follow-up identification and comparison performance.
First aspect, embodiments provides a kind of pedestrian based on multiple gun many balls video camera array
Identification system, described system includes: at least one gunlock, be attached to described gunlock multiple ball machines and
Server;
Described gunlock, for gathering the video image in whole monitoring scene;
Described server, for all moving targets in described video image are carried out detecting and tracking,
And when abnormal motion target being detected, control all ball machines work that described gunlock is corresponding;
Described ball machine, under the control of described server, adjusts the acquisition parameter of self, according to
Acquisition parameter after adjustment gathers the high-resolution of the abnormal motion target that its attached gunlock detects
Video image;Described acquisition parameter at least includes the following parameter of ball machine: level angle, luffing angle
And optical magnification;
Described server is additionally operable to, and extracts the high-resolution of the abnormal motion target that all described ball machines gather
Various dimensions feature in rate video image, and by described various dimensions characteristic storage in corresponding view storehouse
In, in order to the follow-up described various dimensions feature according to storage carries out comparison and the identification of abnormal object;Its
In, described various dimensions feature includes: the association various dimensions feature of face and human body.
In conjunction with first aspect, embodiments provide the first possible embodiment party of first aspect
Formula, wherein, when described gunlock is multiple, sets up the incidence relation between multiple gunlock;
The process of setting up of described incidence relation includes: enter in the gunlock picture that any two gunlock is corresponding
Row scene characteristic Point matching, obtains the feature point set of coupling;Feature point set according to described coupling is set up
Mapping relations between two planes that two gunlock pictures are corresponding;Set up according to described mapping relations and appoint
Target corresponding relation in the scenic picture of two gunlocks of meaning.
In conjunction with the first possible embodiment of first aspect, embodiments provide first party
The embodiment that the second in face is possible, wherein, each described ball machine is at least attached to a gunlock;
When described ball machine is attached to multiple stage gunlock, seizes principle according to high priority and carry out described gunlock to institute
State the control of ball machine;Wherein, the priority of described gunlock is according to following policy setting: the most former
Then or the ball machine visual angle principle of optimality.
In conjunction with the embodiment that the second of first aspect is possible, embodiments provide first party
The third possible embodiment in face, wherein, described server is to owning in described video image
Moving target carries out detecting and tracking and includes:
All moving targets in video image described in described server selection, to described all motion mesh
Target running orbit carries out real-time tracking and analysis, and described running orbit correspondence exception row detected
For time, arranging moving target corresponding to described movement locus is abnormal motion target, and increases described different
The often priority of moving target;Described Deviant Behavior at least includes: quickly run, drive in the wrong direction, be detained and
Fight.
In conjunction with the embodiment that the second of first aspect is possible, embodiments provide first party
The 4th kind of possible embodiment in face, wherein, described server is to owning in described video image
Moving target carries out detecting and tracking, also includes:
All moving targets in video image described in described server selection, according to described moving target
Position in described gunlock picture and the direction of motion, it is judged that whether described moving target will leave
Current scene areas;The most then arranging described moving target is abnormal motion target, and increases institute
State the priority of abnormal motion target.
In conjunction with the third possible embodiment or the 4th kind of possible embodiment of first aspect, this
Inventive embodiments provides the 5th kind of possible embodiment of first aspect, wherein, described server
It is additionally operable to, controls all ball machines corresponding to described gunlock according to abnormal motion target priority height principle
Capture the high-definition picture of described abnormal motion target.
In conjunction with the 5th kind of possible embodiment of first aspect, embodiments provide first party
The 6th kind of possible embodiment in face, wherein, described server is additionally operable to, to all moving targets
It is marked, and the priority arranging the described moving target captured is 0;Wherein, priority is
0 expression is not captured.
In conjunction with first aspect, embodiments provide the 7th kind of possible embodiment party of first aspect
Formula, wherein, ball machine described in described server controls adjusts the acquisition parameter of self and includes:
Described server, when abnormal motion target being detected, positions described abnormal motion target described
Target location in gunlock, arranges ball machine according to described target location and controls parameter, with according to described ball
Machine controls the state modulator ball machine high-resolution video with optimal viewing visual angle acquisition abnormity moving target
Image;Described ball machine controls parameter and includes the following parameter of described ball machine: level angle, luffing angle
And optical magnification.
In conjunction with the 5th kind of possible embodiment of first aspect, embodiments provide first party
The 8th kind of possible embodiment in face, wherein, described server is to owning in described video image
Moving target carries out detecting and tracking, also includes:
Described server uses Faster RCNN algorithm to position all motion mesh in described video image
Mark, after positioning described moving target, uses crowd's dividing method based on degree of depth study to all fortune
Moving-target carries out dividing processing, and uses particle filter tracking algorithm to all motions after dividing processing
The running orbit of target carries out real-time tracking and analysis.
In conjunction with first aspect, embodiments provide the 9th kind of possible embodiment party of first aspect
Formula, wherein, described server extracts the high-resolution of the abnormal motion target that all described ball machines gather
Various dimensions feature in video image includes:
Described server is detecting the high-resolution video figure of abnormal motion target that described ball machine gathers
During picture, catch the face picture in described high resolution video image and the picture of human body, respectively from institute
State face various dimensions feature and the human body various dimensions extracting correspondence in face picture and described human body picture
Feature, obtains the association various dimensions feature of face and human body.
A kind of based on multiple gun many balls video camera array the pedestrian that the embodiment of the present invention provides identifies system,
Including: at least one gunlock, the multiple ball machines being attached to gunlock and server;It is carried out by gunlock
Scene monitoring on a large scale, and gather the abnormal motion target in image by server identification gunlock,
Abnormal motion target in the gunlock that control marble forming machine is attached to it simultaneously is captured, to obtain this exception
The high-definition picture of moving target, the final face extracted in this abnormal motion target and the pass of human body
Connection various dimensions feature preserves, and in order to carry out comparison and the identification of follow-up abnormal object, it is with existing
The resolution having the video image gathered by high definition gunlock in technology is relatively low, when it is used as evidence
In time identifying comparison target, accuracy of identification is poor to be compared, and solves large-range monitoring simultaneously and local is high
The problem of clear monitoring, and for interested or key monitoring target, can be grabbed many by ball machine
The image at individual visual angle, it is possible to guarantee information complete and effectively, improves follow-up identification and comparison performance;
Identify and comparison merit it addition, realize high performance pedestrian by the Feature Fusion mode of face and human body
Can, be also accuracy of identification and accuracy rate higher.
For making the above-mentioned purpose of the present invention, feature and advantage to become apparent, preferable reality cited below particularly
Execute example, and coordinate appended accompanying drawing, be described in detail below.
Accompanying drawing explanation
In order to be illustrated more clearly that the technical scheme of the embodiment of the present invention, below will be to required in embodiment
Accompanying drawing to be used is briefly described, it will be appreciated that the following drawings illustrate only some of the present invention
Embodiment, is therefore not construed as the restriction to scope, for those of ordinary skill in the art,
On the premise of not paying creative work, it is also possible to obtain other relevant accompanying drawings according to these accompanying drawings.
Fig. 1 shows a kind of based on multiple gun many balls video camera array the row that the embodiment of the present invention is provided
People identifies the structural representation of system.
Primary identity illustrates: 101, gunlock;102, ball machine (specifically including 102a, 102b and 102c);
103, server.
Detailed description of the invention
For making the purpose of the embodiment of the present invention, technical scheme and advantage clearer, below in conjunction with this
Accompanying drawing in inventive embodiments, is clearly and completely described the technical scheme in the embodiment of the present invention,
Obviously, described embodiment is only a part of embodiment of the present invention rather than whole embodiments.
Generally herein described in accompanying drawing and the assembly of the embodiment of the present invention that illustrates can be with various different joining
Put and arrange and design.Therefore, retouching in detail the embodiments of the invention provided in the accompanying drawings below
State the scope being not intended to limit claimed invention, but be merely representative of the selected reality of the present invention
Execute example.Based on embodiments of the invention, those skilled in the art are not before making creative work
Put the every other embodiment obtained, broadly fall into the scope of protection of the invention.
In view of traditional face identification system, it is generally individually to use a high definition gunlock to carry out in scene
Video image acquisition, and by server to gather scene video image carry out dynamic human face identification,
Then the analysis result of this video image and corresponding face characteristic is preserved, in order to be subsequently used for
Evidence is laid in investigation and the conduct of event afterwards;And in above-mentioned face identification system, monitoring range little and
Image resolution ratio is looked into, and one of carries out the means of target recognition and comparison separately through face characteristic,
Also make accuracy of identification poor.
Based on the problems referred to above, embodiments provide a kind of based on multiple gun many balls video camera array
Pedestrian identifies system, is described below by embodiment.
The structure of system is identified with reference to a kind of based on multiple gun many balls video camera array the pedestrian shown in Fig. 1
Schematic diagram, illustrate only a gunlock 101 and is attached to three ball machines of above-mentioned gunlock 101 in Fig. 1
102;Embodiments provide a kind of pedestrian based on multiple gun many balls video camera array and identify system,
System includes: at least one gunlock 101, the multiple ball machines 102 being attached to gunlock 101 and server
103;Wherein, each gunlock 101 is all to there being multiple ball machine 102 being erected at diverse location;
Gunlock 101, for gathering the video image in whole monitoring scene;
Server 103, for all moving targets in video image being carried out detecting and tracking, and
When abnormal motion target being detected, all ball machines 102 controlling gunlock 101 correspondence work;
Ball machine 102, under the control of server 103, adjusts the acquisition parameter of self, according to
Acquisition parameter after adjustment gathers the high score of the abnormal motion target that its attached gunlock 101 detects
Resolution video image;Acquisition parameter at least includes the following parameter of ball machine 102: level angle, pitching
Angle and optical magnification;
Server 103 is additionally operable to, and extracts the high-resolution of the abnormal motion target that all ball machines 102 gather
Various dimensions feature in rate video image, and by various dimensions characteristic storage in corresponding view storehouse, with
The continuous various dimensions feature according to storage carries out comparison and the identification of abnormal object after an action of the bowels;Wherein, various dimensions
Feature includes: the association various dimensions feature of face and human body.
A kind of based on multiple gun many balls video camera array the pedestrian that the embodiment of the present invention provides identifies system,
Including: at least one gunlock 101, the multiple ball machines 102 being attached to gunlock 101 and server 103;
It carries out large-scale scene monitoring by gunlock 101, and identifies gunlock 101 by server 103
Gather the abnormal motion target in image, in the gunlock 101 that control marble forming machine 102 is attached to it simultaneously
Abnormal motion target is captured, and to obtain the high-definition picture of this abnormal motion target, finally carries
The association various dimensions feature taking the face in this abnormal motion target and human body preserves, in order to carry out
The comparison of follow-up abnormal object and identification, its with of the prior art by regarding that high definition gunlock gathers
Frequently the resolution of image is relatively low, and when it is used for identifying comparison target as evidence, accuracy of identification is poor
Compare, solve large-range monitoring and the problem of local high-definition monitoring simultaneously, and for interested or
Key monitoring target, can grab the image at multiple visual angle, it is possible to guarantee information by ball machine 102
Complete and effectively, improve follow-up identification and comparison performance;It addition, by face and the spy of human body
Levy amalgamation mode to realize high performance pedestrian and identify and comparison function, be also accuracy of identification and accuracy rate relatively
High.
The many balls of multiple gun in the embodiment of the present invention refer to: multiple gunlock 101 video cameras and multiple ball machine
The video camera array of 102 video camera compositions.
In embodiments of the present invention, first select large-range monitoring scene, to selected large-range monitoring field
Scape, sets up some gunlocks 101 and ball machine 102 in its surrounding, and gunlock 101 is responsible for monitoring whole scene,
It is characterized in that monitoring range is big, but the high-definition picture of local cannot be obtained;Ball machine 102 is responsible for pressing
Acquisition parameter (such as visual angle and optical magnification) need to be adjusted, and according to according to the collection ginseng after adjusting
The high resolution video image of moving target in number photographed scene.
Server 103 then moves in gunlock picture (video image that i.e. gunlock 101 gathers)
Target carries out detecting and tracking, after finding the target (based on pedestrian) merited attention, controls gunlock
101 attached all ball machines 102 carry out multi-faceted high-resolution candid photograph, it is thus achieved that the high-resolution of this target
Rate video segment, and the various dimensions feature in this high-definition picture is extracted by Intellectual Analysis Technology.
This various dimensions feature will be used for succeeding target identification and comparison.Moving target in the embodiment of the present invention can
Think pedestrian, vehicle or animal etc., the motion that the identification system in the embodiment of the present invention is mainly applied
Target is pedestrian.
Infrastructure from identification system carries out whole in terms of building and identify the operation two of system below
Body is introduced:
First, the infrastructure in above-mentioned identification system is built and is briefly described:
1, monitoring scene is first selected: the identification system that the embodiment of the present invention provides applies in general to monitoring
Scene is relatively big, and high-altitude is blocked little, and surrounding has the situation of high constructure.
2, the some position of gunlock 101 sets up: during actual monitored, gunlock 101 can be one,
Can also be multiple stage, and the determination of amount of gunlock 101, as long as the visual field meeting gunlock 101 is permissible
Cover whole monitoring scene, and gunlock 101 at decorating position preferably in higher position.
3,102 positions of ball machine set up: ball machine 102 is attached to above-mentioned gunlock 101, ball machine 102
It is generally required at least three billiard ball machines 102 (being recommended as four), it is therefore an objective to ball machine 102 is distributed erection
At the four direction of monitoring scene, so that ball machine 102 is distributed as evenly as possible monitoring scene four
Week;In the embodiment of the present invention, ball machine 102 antenna height selects: the motion in ball machine 102 and scene
The angle of target link and ground level are more preferably less than 30 degree, and in this, erection mode is conducive to ball machine 102
Get optimal face.
4, network service can be realized between gunlock 101, ball machine 102 and server 103.
In the embodiment of the present invention, gunlock 101 and ball machine 102 can be all-in-ones.Above-mentioned server 103
Can be intellectual analysis server, be arranged in the rear end of gunlock 101, with respectively to gunlock 101 and ball
The video image that machine 102 gathers processes accordingly;Above-mentioned server 103 can also be embedded
Intelligent analysis module, be embedded within gunlock 101, respectively gunlock 101 and ball machine 102 are adopted
The video image of collection processes accordingly;Above-mentioned corresponding process includes: server 103 is to gunlock
101 video images gathered carry out detecting and tracking, it is also possible to be controlled all ball machines 102, with
And the high-resolution video image of ball machine 102 collection is carried out feature extraction and preserves.
In the embodiment of the present invention, large scene monitors: the monitoring scene that generally referred to as scope is bigger,
Video camera distance detection target is distant, and such as square monitoring, harbour monitor, are placed in mansion roof
Ground monitoring etc..Pedestrian identifies: can identify the identity of people, or automatically judge two in multiple scenes
Whether individual is same person, also referred to as Person Reidentification (i.e. Authenticate afresh or
Person re-recognizes) technology.
Second, the process of the operation of above-mentioned identification system is described in detail:
1, first carry out system calibrating: as above, in the embodiment of the present invention, gunlock 101 can one,
Can also be multiple, when gunlock 101 is multiple, set up the incidence relation between multiple gunlock 101;
The process of setting up of incidence relation includes: carry out field in the gunlock picture of any two gunlock 101 correspondence
Scape Feature Points Matching, obtains the feature point set of coupling;Feature point set according to coupling sets up two gunlocks
Mapping relations between two planes that picture is corresponding;Any two gunlock is set up according to mapping relations
Target corresponding relation in the scenic picture of 101.
Concrete, when gunlock 101 is multiple, set up the corresponding relation between gunlock 101: if
Visual field between two gunlocks 101 does not overlap, then it is assumed that between two gunlocks 101, it doesn't matter,
This demarcation mode is the most common;Demarcation mode i.e. can be passed through in the way of using above-mentioned automatic Calibration
Ground level is assumed to demarcate the relation between two gunlocks 101, specifically according to traditional computer vision
Method, first directly carries out scene characteristic Point matching in two gunlock pictures, according to the feature of coupling
Point set can set up the mapping relations between two planes, thus sets up the target pair in two scenic picture
Should be related to.Assume that method demarcates the relation between two gunlocks 101 by above-mentioned ground level, it is therefore an objective to
Ensure that the target corresponding relation in two scenic picture is linear;Generally, above-mentioned calibration process is permissible
Realized by the scaling method of 4 points, i.e. select four scene characteristic points to mate.On having had
After the demarcation relation stated, certain target seen in a gunlock 101, can directly correspond to another
The position at same target place in the picture of one gunlock 101.
2, gunlock 101 and the direct corresponding relation of ball machine 102 are set up: each ball machine 102 is the most attached
Belong to a gunlock 101;When ball machine 102 is attached to multiple stage gunlock 101, rob according to high priority
Account for principle and carry out the gunlock 101 control to ball machine 102;Wherein, the priority of gunlock 101 is basis
Following policy setting: nearby principle or the ball machine 102 visual angle principle of optimality.
Concrete, gunlock 101 and the direct corresponding relation of ball machine 102 belong to the rifle ball of standard and demarcate,
Rifle ball is demarcated and can be provided by Qiang Qiu equipment vendors, and some manufacturers provide automatic Calibration, i.e. use ibid
The method stating the Feature Points Matching that the foundation of the corresponding relation between multiple gunlock 101 is used, some factories
Business provides manual scaling method: i.e. by the point of artificial selected gunlock picture, (this point is it is generally required at least
3 points), then adjust ball machine 102 visual angle and make this point be in lean-with in ball machine 102 picture
Put, to complete to demarcate.
Further, each ball machine 102 is at least attached to a gunlock 101, if a ball machine 102 belongs to
In multiple stage gunlock 101, each gunlock 101 has priority difference to the control of ball machine 102.When
When multiple stage gunlock 101 controls a billiard ball machine 102 simultaneously, it then follows high priority seizes principle, and gunlock
The setting of the priority of 101 can be typically nearby principle, or the most former according to ball machine 102 visual angle
Then.Wherein, nearby principle is i.e.: gunlock 101 priority nearest apart from this ball machine 102 is the highest;Ball
The machine 102 visual angle principle of optimality is i.e.: for a target in gunlock 101, if the ball machine of association
102 can see this target by the control break visual angle of gunlock 101, such as, define target sizes
For index, the target seen by changing visual angle when ball machine 102 pixels tall in picture is the biggest,
Then the visual angle of ball machine 102 is the best, corresponding, control the gunlock 101 of the best ball machine 102 in this visual angle
Priority the highest;The priority of the gunlock 101 of the final ball machine 102 controlling optimal viewing angle is the highest.
3, demarcate the target location in gunlock 101 and ball machine 102 control the relation of parameter:
Server 103 is controlled marble forming machine 102 and is adjusted self acquisition parameter and include: server 103 is in inspection
When measuring abnormal motion target, position abnormal motion target target location in gunlock 101, according to
Target location arranges ball machine 102 and controls parameter, to control state modulator ball machine 102 according to ball machine 102
High resolution video image with optimal viewing visual angle acquisition abnormity moving target;Ball machine 102 controls ginseng
Several following parameters including ball machine 102: level angle, luffing angle and optical magnification.
Concrete, ball machine 102 controls the relation of parameter, including level angle, the pitching of ball machine 102
The 3D location of angle and optical magnification, also referred to as ball machine 102.Target in gunlock 101
Position and ball machine 102 control the corresponding relation of parameter primarily to obtain in ball machine 102 visual field
Optimal viewing visual angle, is i.e. determined by level angle and the luffing angle of ball machine 102, make shooting (or
Person says candid photograph) target the most placed in the middle;It is determined by the optical magnification of ball machine 102, makes bat
The picture complete packet taking the photograph (capturing in other words) contains target, and target detail sees clearly as far as possible.
In reality, owing to target may be moved, first the position of target in gunlock 101 is amplified,
Then ball machine 102 is set according to the target location after amplifying and controls parameter, it is to avoid target moves out ball machine
102 visuals field.
In the embodiment of the present invention, all moving targets in video image can also be entered by server 103
Row detecting and tracking, concrete grammar includes:
All moving targets in server 103 positioning video image, the operation to all moving targets
Track carries out real-time tracking and analysis, and when running orbit correspondence Deviant Behavior being detected, arranges fortune
The moving target that dynamic track is corresponding is abnormal motion target, and increases the priority of abnormal motion target;
Deviant Behavior at least includes: quickly runs, drive in the wrong direction, be detained and fight;
All moving targets in server 103 positioning video image, draw at gunlock according to moving target
Position in face and the direction of motion, it is judged that whether moving target will leave current scene areas,
The most then arranging moving target is abnormal motion target, and increases the priority of abnormal motion target.
Concrete, the method for all moving targets in server 103 positioning video image specifically:
Server 103 uses Faster RCNN algorithm to carry out human detection, to position pedestrian, and at clothes
After all moving targets (such as pedestrian) in business device 103 positioning video image, use based on the degree of depth
The crowd's dividing method practised carries out crowd's dividing processing to all moving targets (such as pedestrian), and uses
Particle filter tracking algorithm the running orbit of all moving targets after dividing processing is carried out in real time with
Track and analysis.
All of moving target in the video image that gunlock 101 is gathered by server 103 has carried out excellent
First level divides, and obtains the moving target according to priority arrangement, and correspondence, server 103 then controls
All ball machines 102 of gunlock 101 correspondence are captured abnormal according to abnormal motion target priority height principle
The high-definition picture of moving target.
Server 103 control correspondence ball machine 102 captured abnormal motion target while, service
All moving targets are marked, and arrange the priority of the moving target captured in real time by device 103
It is 0;Wherein, priority is that 0 expression is not captured.
In the embodiment of the present invention, the video image that all gunlocks 101 are gathered by server 103 enters in real time
Row moving object detection and tracking, is split by human detection auxiliary crowd simultaneously.Concrete, service
The video image that all gunlocks 101 are gathered by device 103 by Faster RCNN algorithm carries out people in real time
Health check-up is surveyed, and to position pedestrian, uses crowd's dividing method based on degree of depth study to carry out crowd simultaneously and divides
Cut, with the crowd that flocks together of differentiation, prevent the crowd flocked together as a moving target,
And by particle filter tracking algorithm, the movement locus of the human body after crowd is split carry out in real time with
Track.
Server 103 is after carrying out real-time tracking to the movement locus of pedestrian, to all moving targets (i.e.
Pedestrian) it is made to determine whether to capture its high-resolution features, and according to all motions of following policy setting
The priority of target (i.e. pedestrian), wherein, the default priority of moving target is 1 (basic value),
0 expression is not captured;When the movement locus that server 103 detects human body is Deviant Behavior,
Increase the candid photograph priority of moving target corresponding to this movement locus (i.e. pedestrian), i.e. excellent give tacit consent to
Increase default value on the basis of first level basic value, and the numerical value specifically increased can be according to the fortune of human body
Dynamic track determines with the similarity of Deviant Behavior, as added 1, can add 2.
If the attached ball machine 102 of gunlock 101 needs to capture multiple moving target, then server 103
Controlling corresponding ball machine 102 to capture according to the shooting priority of moving target, i.e. preferential shooting is worked as
The moving target (i.e. pedestrian) that front priority is the highest.If all moving targets (i.e. pedestrian) is excellent
First level is all 0, then server 103 is controlled marble forming machine 102 and captured at random.
All pursuit movement targets (i.e. pedestrian) can be marked, for having grabbed by server 103
Clapping the moving target (i.e. pedestrian) of high-resolution features, directly arranging priority is 0, no longer captures.
Server 103 carries out Deviant Behavior judgement to the running orbit of moving target, when operation being detected
When track corresponds to Deviant Behavior, increase the priority of moving target corresponding to this running orbit (as excellent
First level adds 1), with the preferential candid photograph to abnormal motion target;These include extremely: quickly run,
Drive in the wrong direction, be detained for a long time, fight.
Server 103 is according to moving target position in gunlock picture and the direction of motion, it is judged that mesh
Whether mark will leave visual field, increase the priority of moving target at the scene left (as preferentially simultaneously
Level adds 1), preferentially capture so that the moving target of visual field will be left.
Ball machine 102 performs high definition snapshot task under the control of server 103: when certain moving target
After judging to need to carry out high resoluting characteristic extraction, according to the gunlock 101 belonging to this moving target, control
This target dynamically paid close attention to by all ball machines 102 of this gunlock 101 subordinate, and in order to ensure that each pass is gazed at
Mark can capture the high-resolution features of various visual angles as far as possible, and synchronization only captures a target, institute
With, it can be seen that the ball machine 102 of this target, all can perform the candid photograph task of this target.Server 103
Ball machine 102 video can be analyzed in real time, owing to now target is the most amplified, face can be carried out
Detection, tracking and feature extraction with human body, these information will be saved in backstage view storehouse, with
Just it is subsequently used for carrying out the retrieval analysis of face and human body.
Concrete, each high-resolution of ball machine 102 captures task default time, is traditionally arranged to be 3-5
Second, i.e. ball machine 102 starts to change visual angle after receiving candid photograph task and optical magnification is grabbed
Clap, i.e. terminate after the 3-5 second to capture task specifically.
Further, the identification system that the present invention provides supports a region in the most selected scene, arranges
For limit priority, other being carrying out of tasks can be interrupted.Relevant ball machine 102 carries out high definition
Monitoring, the monitoring period of all ball machines 102 (capture and appoint by ball machine 102 high-resolution the most above-mentioned
The business time) it is defaulted as 10 seconds, this time can configure the most as required, and ball machine 102 is complete
During after one-tenth, high-definition monitoring task automatically returns to normal monitoring flow process.When ball machine 102 is the most more high
During clear candid photograph task, it returns to any one preset position, as commonly monitoring use.
In embodiments of the present invention, server 103 extracts the abnormal motion mesh that all ball machines 102 gather
Various dimensions feature in target high resolution video image includes:
Server 103 is detecting the high-resolution video figure of abnormal motion target that ball machine 102 gathers
During picture, catch the face picture in high resolution video image and the picture of human body, draw from face respectively
Face and human body picture are extracted face various dimensions feature and the human body various dimensions feature of correspondence, obtains face
Association various dimensions feature with human body.
Concrete, shoot by 102 degree of abnormal motion targets of multiple ball machines or capture, to same
One target whole identification system can capture various visual angles, the details video image of many time points, by this
A little images, server 103 can extract the face of moving target (i.e. pedestrian) of correspondence, human body
Deng various dimensions information.The comparison with abnormal object and identification, recognition of face skill is carried out for the ease of follow-up
Art require positive face, and the embodiment of the present invention provide system, i.e. person who happens to be on hand for an errand towards uncontrolled, regarding more
Under angle is captured, the biggest chance is also had to catch several positive faces;And multi-angle picture, can carry equally
Take more complete feature, be conducive to carrying out human body retrieval;The such as back side and front clothing feature difference
The biggest situation, has very great help for improving retrieval success rate.Therefore, server 103 can be automatically
Screen and associate more features to same target, form one group of effective face, characteristics of human body's collection
Close, and therefrom extract some structured attributes, include whether wearing spectacles, whether be branded as, upper body
Color, lower part of the body color etc..These information can be used for carrying out comparison and the identification of abnormal object, such as view
The text retrieval in storehouse, recognition of face, human body retrieval etc..
The pedestrian based on multiple gun many balls video camera array that the embodiment of the present invention provides identifies system, permissible
Apply in multiple scenes, as being used in square, for unlawful assembly, or to paying close attention to people
Member deploys to ensure effective monitoring and control of illegal activities;Government unit doorway can be used in, for person appealing for help, carry out identification, pre-
Alert;Can be used in market, for identification, the keypoint recommendation of client.
A kind of based on multiple gun many balls video camera array the pedestrian that the embodiment of the present invention provides identifies system,
Including: at least one gunlock 101, the multiple ball machines 102 being attached to gunlock 101 and server 103;
It carries out large-scale scene monitoring by gunlock 101, and identifies gunlock 101 by server 103
Gather the abnormal motion target in image, in the gunlock 101 that control marble forming machine 102 is attached to it simultaneously
Abnormal motion target is captured, and to obtain the high-definition picture of this abnormal motion target, finally carries
The association various dimensions feature taking the face in this abnormal motion target and human body preserves, in order to carry out
The comparison of follow-up abnormal object and identification, its with of the prior art by regarding that high definition gunlock gathers
Frequently the resolution of image is relatively low, and when it is used for identifying comparison target as evidence, accuracy of identification is poor
Compare, solve large-range monitoring and the problem of local high-definition monitoring simultaneously, and for interested or
Key monitoring target, can grab the image at multiple visual angle, it is possible to guarantee information by ball machine 102
Complete and effectively, improve follow-up identification and comparison performance;It addition, by face and the spy of human body
Levy amalgamation mode to realize high performance pedestrian and identify and comparison function, be also accuracy of identification and accuracy rate relatively
High.
It should also be noted that similar label and letter represent similar terms, therefore, one in following accompanying drawing
A certain Xiang Yi the accompanying drawing of denier is defined, then need not it is carried out further in accompanying drawing subsequently
Definition and explanation, describe additionally, term " first ", " second ", " the 3rd " etc. are only used for distinguishing,
And it is not intended that indicate or hint relative importance.
Last it is noted that the detailed description of the invention of embodiment described above, the only present invention, use
So that technical scheme to be described, being not intended to limit, protection scope of the present invention is not limited to
This, is although the present invention being described in detail with reference to previous embodiment, the ordinary skill of this area
Personnel are it is understood that any those familiar with the art is at the technical scope that the invention discloses
In, the technical scheme described in previous embodiment still can be modified by it maybe can readily occur in change
Change, or wherein portion of techniques feature is carried out equivalent;And these are revised, change or replace,
The essence not making appropriate technical solution departs from the spirit and scope of embodiment of the present invention technical scheme.All
Should contain within protection scope of the present invention.Therefore, protection scope of the present invention should be described with right
The protection domain required is as the criterion.
Claims (10)
1. a pedestrian based on multiple gun many balls video camera array identifies system, it is characterised in that described
System includes: at least one gunlock, the multiple ball machines being attached to described gunlock and server;
Described gunlock, for gathering the video image in whole monitoring scene;
Described server, for all moving targets in described video image are carried out detecting and tracking,
And when abnormal motion target being detected, control all ball machines work that described gunlock is corresponding;
Described ball machine, under the control of described server, adjusts the acquisition parameter of self, according to
Acquisition parameter after adjustment gathers the high-resolution of the abnormal motion target that its attached gunlock detects
Video image;Described acquisition parameter at least includes the following parameter of ball machine: level angle, luffing angle
And optical magnification;
Described server is additionally operable to, and extracts the high-resolution of the abnormal motion target that all described ball machines gather
Various dimensions feature in rate video image, and by described various dimensions characteristic storage in corresponding view storehouse
In, in order to the follow-up described various dimensions feature according to storage carries out comparison and the identification of abnormal object;Its
In, described various dimensions feature includes: the association various dimensions feature of face and human body.
Pedestrian based on multiple gun many balls video camera array the most according to claim 1 identifies system,
It is characterized in that, when described gunlock is multiple, set up the incidence relation between multiple gunlock;
The process of setting up of described incidence relation includes: enter in the gunlock picture that any two gunlock is corresponding
Row scene characteristic Point matching, obtains the feature point set of coupling;Feature point set according to described coupling is set up
Mapping relations between two planes that two gunlock pictures are corresponding;Set up according to described mapping relations and appoint
Target corresponding relation in the scenic picture of two gunlocks of meaning.
Pedestrian based on multiple gun many balls video camera array the most according to claim 2 identifies system,
It is characterized in that, each described ball machine is at least attached to a gunlock;When described ball machine is attached to multiple stage
During gunlock, seize principle according to high priority and carry out the control to described ball machine of the described gunlock;Wherein,
The priority of described gunlock is according to following policy setting: nearby principle or ball machine visual angle are the most former
Then.
Pedestrian based on multiple gun many balls video camera array the most according to claim 3 identifies system,
It is characterized in that, described server carries out detecting and tracking to all moving targets in described video image
Including:
All moving targets in video image described in described server selection, to described all motion mesh
Target running orbit carries out real-time tracking and analysis, and described running orbit correspondence exception row detected
For time, arranging moving target corresponding to described movement locus is abnormal motion target, and increases described different
The often priority of moving target;Described Deviant Behavior at least includes: quickly run, drive in the wrong direction, be detained and
Fight.
Pedestrian based on multiple gun many balls video camera array the most according to claim 3 identifies system,
It is characterized in that, described server all moving targets in described video image are detected with
Track, also includes:
All moving targets in video image described in described server selection, according to described moving target
Position in described gunlock picture and the direction of motion, it is judged that whether described moving target will leave
Current scene areas;The most then arranging described moving target is abnormal motion target, and increases institute
State the priority of abnormal motion target.
6. identify system according to the pedestrian based on multiple gun many balls video camera array described in claim 4 or 5
System, it is characterised in that described server is additionally operable to, control all ball machines corresponding to described gunlock according to
Abnormal motion target priority height principle captures the high-definition picture of described abnormal motion target.
Pedestrian based on multiple gun many balls video camera array the most according to claim 6 identifies system,
It is characterized in that, described server is additionally operable to, and is marked all moving targets, and arranges candid photograph
The priority of the described moving target crossed is 0;Wherein, priority is that 0 expression is not captured.
Pedestrian based on multiple gun many balls video camera array the most according to claim 1 identifies system,
It is characterized in that, ball machine described in described server controls adjusts the acquisition parameter of self and includes:
Described server, when abnormal motion target being detected, positions described abnormal motion target described
Target location in gunlock, arranges ball machine according to described target location and controls parameter, with according to described ball
Machine controls the state modulator ball machine high-resolution video with optimal viewing visual angle acquisition abnormity moving target
Image;Described ball machine controls parameter and includes the following parameter of described ball machine: level angle, luffing angle
And optical magnification.
Pedestrian based on multiple gun many balls video camera array the most according to claim 6 identifies system,
It is characterized in that, described server all moving targets in described video image are detected with
Track, also includes:
Described server uses Faster RCNN algorithm to position all motion mesh in described video image
Mark, after positioning described moving target, uses crowd's dividing method based on degree of depth study to all fortune
Moving-target carries out dividing processing, and uses particle filter tracking algorithm to all motions after dividing processing
The running orbit of target carries out real-time tracking and analysis.
Pedestrian based on multiple gun many balls video camera array the most according to claim 1 identifies system,
It is characterized in that, described server extracts the high-resolution of the abnormal motion target that all described ball machines gather
Various dimensions feature in rate video image includes:
Described server is detecting the high-resolution video figure of abnormal motion target that described ball machine gathers
During picture, catch the face picture in described high resolution video image and the picture of human body, respectively from institute
State face various dimensions feature and the human body various dimensions extracting correspondence in face picture and described human body picture
Feature, obtains the association various dimensions feature of face and human body.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610396862.XA CN105979210B (en) | 2016-06-06 | 2016-06-06 | A kind of pedestrian's identifying system based on the more ball video camera arrays of multiple gun |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610396862.XA CN105979210B (en) | 2016-06-06 | 2016-06-06 | A kind of pedestrian's identifying system based on the more ball video camera arrays of multiple gun |
Publications (2)
Publication Number | Publication Date |
---|---|
CN105979210A true CN105979210A (en) | 2016-09-28 |
CN105979210B CN105979210B (en) | 2019-02-05 |
Family
ID=57010230
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610396862.XA Expired - Fee Related CN105979210B (en) | 2016-06-06 | 2016-06-06 | A kind of pedestrian's identifying system based on the more ball video camera arrays of multiple gun |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN105979210B (en) |
Cited By (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN106791706A (en) * | 2017-01-24 | 2017-05-31 | 上海木爷机器人技术有限公司 | Object lock method and system |
CN107845105A (en) * | 2017-10-24 | 2018-03-27 | 深圳市圆周率软件科技有限责任公司 | A kind of monitoring method, smart machine and storage medium based on the linkage of panorama rifle ball |
CN108093209A (en) * | 2016-11-21 | 2018-05-29 | 松下知识产权经营株式会社 | Image transmission system and dollying machine equipment |
CN108108662A (en) * | 2017-11-24 | 2018-06-01 | 深圳市华尊科技股份有限公司 | Deep neural network identification model and recognition methods |
CN108243304A (en) * | 2016-12-23 | 2018-07-03 | 厦门博聪信息技术有限公司 | Rifle ball integral type dynamic human face captures dedicated video camera and dynamic human face grasp shoot method |
CN108256401A (en) * | 2016-12-29 | 2018-07-06 | 杭州海康威视数字技术股份有限公司 | A kind of method and device for obtaining objective attribute target attribute Feature Semantics |
CN108521554A (en) * | 2018-03-01 | 2018-09-11 | 西安电子科技大学 | Large scene multi-target cooperative tracking method, intelligent monitor system, traffic system |
CN109543645A (en) * | 2018-11-30 | 2019-03-29 | 北京小马智行科技有限公司 | A kind of face identification method and device |
CN109711369A (en) * | 2018-12-29 | 2019-05-03 | 深圳英飞拓智能技术有限公司 | Pedestrian count method, apparatus, system, computer equipment and storage medium |
CN109886730A (en) * | 2019-01-15 | 2019-06-14 | 北京奇艺世纪科技有限公司 | Put-on method, equipment and system, service equipment and method, processing equipment, medium |
CN109922250A (en) * | 2017-12-12 | 2019-06-21 | 杭州海康威视数字技术股份有限公司 | A kind of target object grasp shoot method, device and video monitoring equipment |
CN110175535A (en) * | 2019-05-08 | 2019-08-27 | 广州中交通信有限公司 | A kind of ship identification scheme and its recognition methods based on deep learning |
CN110706247A (en) * | 2018-06-22 | 2020-01-17 | 杭州海康威视数字技术股份有限公司 | Target tracking method, device and system |
CN110930716A (en) * | 2019-11-12 | 2020-03-27 | 浙江宇视科技有限公司 | License plate recognition method and device based on cooperation of multiple cameras |
CN110969657A (en) * | 2018-09-29 | 2020-04-07 | 杭州海康威视数字技术股份有限公司 | Gun and ball coordinate association method and device, electronic equipment and storage medium |
CN111083358A (en) * | 2019-12-05 | 2020-04-28 | 浙江大华技术股份有限公司 | Target snapshot method, device and system |
WO2020135106A1 (en) * | 2018-12-28 | 2020-07-02 | 杭州海康威视数字技术股份有限公司 | Method and apparatus for tracking target object, monitoring device, and storage medium |
CN111402304A (en) * | 2020-03-23 | 2020-07-10 | 浙江大华技术股份有限公司 | Target object tracking method and device and network video recording equipment |
CN111800605A (en) * | 2020-06-15 | 2020-10-20 | 深圳英飞拓科技股份有限公司 | Gun-ball linkage based vehicle shape and license plate transmission method, system and equipment |
CN111800604A (en) * | 2020-06-12 | 2020-10-20 | 深圳英飞拓科技股份有限公司 | Method and device for detecting human shape and human face data based on gun and ball linkage |
CN111914592A (en) * | 2019-05-08 | 2020-11-10 | 杭州海康威视数字技术股份有限公司 | Multi-camera combined evidence obtaining method, device and system |
CN111935403A (en) * | 2020-08-14 | 2020-11-13 | 昆山同日工业自动化有限公司 | Confidential area violation photographing monitoring system based on 5G network |
CN112016483A (en) * | 2020-08-31 | 2020-12-01 | 杭州海康威视数字技术股份有限公司 | Relay system, method, device and equipment for target detection |
CN112488068A (en) * | 2020-12-21 | 2021-03-12 | 重庆紫光华山智安科技有限公司 | Method, device and equipment for searching monitoring target and computer storage medium |
WO2022134916A1 (en) * | 2020-12-24 | 2022-06-30 | 中兴通讯股份有限公司 | Identity feature generation method and device, and storage medium |
CN115379165A (en) * | 2022-07-22 | 2022-11-22 | 交通运输部路网监测与应急处置中心 | Tunnel monitoring method and system based on gunlock and dome camera linkage |
CN116456061A (en) * | 2023-06-15 | 2023-07-18 | 四川三思德科技有限公司 | Intelligent community monitoring management method, system and medium based on dynamic target detection |
CN117278837A (en) * | 2023-11-16 | 2023-12-22 | 新乡天辅电子科技有限公司 | Emergency rescue-oriented imaging equipment control method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102946528A (en) * | 2012-12-14 | 2013-02-27 | 安徽水天信息科技有限公司 | Airport runway monitoring system based on intelligent video monitoring for whole scenic spot |
US20140023357A1 (en) * | 2012-07-18 | 2014-01-23 | Pelco, Inc. | Camera |
CN103761514A (en) * | 2014-01-26 | 2014-04-30 | 公安部第三研究所 | System and method for achieving face recognition based on wide-angle gun camera and multiple dome cameras |
CN103888731A (en) * | 2014-03-24 | 2014-06-25 | 公安部第三研究所 | Structured description device and system for mixed video monitoring by means of gun-type camera and dome camera |
CN104125433A (en) * | 2014-07-30 | 2014-10-29 | 西安冉科信息技术有限公司 | Moving object video surveillance method based on multi-PTZ (pan-tilt-zoom)-camera linkage structure |
CN104639917A (en) * | 2015-03-05 | 2015-05-20 | 苏州科达科技股份有限公司 | Video monitoring system and control method thereof |
CN104902236A (en) * | 2015-05-27 | 2015-09-09 | 深圳英飞拓科技股份有限公司 | Linkage control method and device for monitoring equipment |
-
2016
- 2016-06-06 CN CN201610396862.XA patent/CN105979210B/en not_active Expired - Fee Related
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20140023357A1 (en) * | 2012-07-18 | 2014-01-23 | Pelco, Inc. | Camera |
CN102946528A (en) * | 2012-12-14 | 2013-02-27 | 安徽水天信息科技有限公司 | Airport runway monitoring system based on intelligent video monitoring for whole scenic spot |
CN103761514A (en) * | 2014-01-26 | 2014-04-30 | 公安部第三研究所 | System and method for achieving face recognition based on wide-angle gun camera and multiple dome cameras |
CN103888731A (en) * | 2014-03-24 | 2014-06-25 | 公安部第三研究所 | Structured description device and system for mixed video monitoring by means of gun-type camera and dome camera |
CN104125433A (en) * | 2014-07-30 | 2014-10-29 | 西安冉科信息技术有限公司 | Moving object video surveillance method based on multi-PTZ (pan-tilt-zoom)-camera linkage structure |
CN104639917A (en) * | 2015-03-05 | 2015-05-20 | 苏州科达科技股份有限公司 | Video monitoring system and control method thereof |
CN104902236A (en) * | 2015-05-27 | 2015-09-09 | 深圳英飞拓科技股份有限公司 | Linkage control method and device for monitoring equipment |
Cited By (42)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN108093209A (en) * | 2016-11-21 | 2018-05-29 | 松下知识产权经营株式会社 | Image transmission system and dollying machine equipment |
CN108093209B (en) * | 2016-11-21 | 2024-04-12 | 松下知识产权经营株式会社 | Image transmission system and mobile camera device |
CN108243304A (en) * | 2016-12-23 | 2018-07-03 | 厦门博聪信息技术有限公司 | Rifle ball integral type dynamic human face captures dedicated video camera and dynamic human face grasp shoot method |
CN108256401A (en) * | 2016-12-29 | 2018-07-06 | 杭州海康威视数字技术股份有限公司 | A kind of method and device for obtaining objective attribute target attribute Feature Semantics |
CN108256401B (en) * | 2016-12-29 | 2021-03-26 | 杭州海康威视数字技术股份有限公司 | Method and device for obtaining target attribute feature semantics |
CN106791706A (en) * | 2017-01-24 | 2017-05-31 | 上海木爷机器人技术有限公司 | Object lock method and system |
CN107845105A (en) * | 2017-10-24 | 2018-03-27 | 深圳市圆周率软件科技有限责任公司 | A kind of monitoring method, smart machine and storage medium based on the linkage of panorama rifle ball |
CN107845105B (en) * | 2017-10-24 | 2021-09-10 | 深圳市圆周率软件科技有限责任公司 | Monitoring method based on panoramic gun-ball linkage, intelligent device and storage medium |
CN108108662A (en) * | 2017-11-24 | 2018-06-01 | 深圳市华尊科技股份有限公司 | Deep neural network identification model and recognition methods |
CN108108662B (en) * | 2017-11-24 | 2021-05-25 | 深圳市华尊科技股份有限公司 | Deep neural network recognition model and recognition method |
CN109922250A (en) * | 2017-12-12 | 2019-06-21 | 杭州海康威视数字技术股份有限公司 | A kind of target object grasp shoot method, device and video monitoring equipment |
CN109922250B (en) * | 2017-12-12 | 2021-04-02 | 杭州海康威视数字技术股份有限公司 | Target object snapshot method and device and video monitoring equipment |
CN108521554A (en) * | 2018-03-01 | 2018-09-11 | 西安电子科技大学 | Large scene multi-target cooperative tracking method, intelligent monitor system, traffic system |
CN108521554B (en) * | 2018-03-01 | 2021-03-26 | 西安电子科技大学 | Large-scene multi-target cooperative tracking method, intelligent monitoring system and traffic system |
CN110706247A (en) * | 2018-06-22 | 2020-01-17 | 杭州海康威视数字技术股份有限公司 | Target tracking method, device and system |
CN110969657A (en) * | 2018-09-29 | 2020-04-07 | 杭州海康威视数字技术股份有限公司 | Gun and ball coordinate association method and device, electronic equipment and storage medium |
CN110969657B (en) * | 2018-09-29 | 2023-11-03 | 杭州海康威视数字技术股份有限公司 | Gun ball coordinate association method and device, electronic equipment and storage medium |
CN109543645A (en) * | 2018-11-30 | 2019-03-29 | 北京小马智行科技有限公司 | A kind of face identification method and device |
WO2020135106A1 (en) * | 2018-12-28 | 2020-07-02 | 杭州海康威视数字技术股份有限公司 | Method and apparatus for tracking target object, monitoring device, and storage medium |
CN111383251A (en) * | 2018-12-28 | 2020-07-07 | 杭州海康威视数字技术股份有限公司 | Method and device for tracking target object, monitoring equipment and storage medium |
CN109711369A (en) * | 2018-12-29 | 2019-05-03 | 深圳英飞拓智能技术有限公司 | Pedestrian count method, apparatus, system, computer equipment and storage medium |
CN109886730A (en) * | 2019-01-15 | 2019-06-14 | 北京奇艺世纪科技有限公司 | Put-on method, equipment and system, service equipment and method, processing equipment, medium |
CN110175535B (en) * | 2019-05-08 | 2023-07-04 | 广州中交通信有限公司 | Ship identification system and method based on deep learning |
CN111914592B (en) * | 2019-05-08 | 2023-09-05 | 杭州海康威视数字技术股份有限公司 | Multi-camera combined evidence obtaining method, device and system |
CN111914592A (en) * | 2019-05-08 | 2020-11-10 | 杭州海康威视数字技术股份有限公司 | Multi-camera combined evidence obtaining method, device and system |
CN110175535A (en) * | 2019-05-08 | 2019-08-27 | 广州中交通信有限公司 | A kind of ship identification scheme and its recognition methods based on deep learning |
CN110930716B (en) * | 2019-11-12 | 2021-08-31 | 浙江宇视科技有限公司 | License plate recognition method and device based on cooperation of multiple cameras |
CN110930716A (en) * | 2019-11-12 | 2020-03-27 | 浙江宇视科技有限公司 | License plate recognition method and device based on cooperation of multiple cameras |
CN111083358A (en) * | 2019-12-05 | 2020-04-28 | 浙江大华技术股份有限公司 | Target snapshot method, device and system |
CN111402304A (en) * | 2020-03-23 | 2020-07-10 | 浙江大华技术股份有限公司 | Target object tracking method and device and network video recording equipment |
CN111800604A (en) * | 2020-06-12 | 2020-10-20 | 深圳英飞拓科技股份有限公司 | Method and device for detecting human shape and human face data based on gun and ball linkage |
CN111800605A (en) * | 2020-06-15 | 2020-10-20 | 深圳英飞拓科技股份有限公司 | Gun-ball linkage based vehicle shape and license plate transmission method, system and equipment |
CN111935403A (en) * | 2020-08-14 | 2020-11-13 | 昆山同日工业自动化有限公司 | Confidential area violation photographing monitoring system based on 5G network |
CN112016483A (en) * | 2020-08-31 | 2020-12-01 | 杭州海康威视数字技术股份有限公司 | Relay system, method, device and equipment for target detection |
CN112488068A (en) * | 2020-12-21 | 2021-03-12 | 重庆紫光华山智安科技有限公司 | Method, device and equipment for searching monitoring target and computer storage medium |
CN112488068B (en) * | 2020-12-21 | 2022-01-11 | 重庆紫光华山智安科技有限公司 | Method, device and equipment for searching monitoring target and computer storage medium |
WO2022134916A1 (en) * | 2020-12-24 | 2022-06-30 | 中兴通讯股份有限公司 | Identity feature generation method and device, and storage medium |
CN115379165A (en) * | 2022-07-22 | 2022-11-22 | 交通运输部路网监测与应急处置中心 | Tunnel monitoring method and system based on gunlock and dome camera linkage |
CN116456061A (en) * | 2023-06-15 | 2023-07-18 | 四川三思德科技有限公司 | Intelligent community monitoring management method, system and medium based on dynamic target detection |
CN116456061B (en) * | 2023-06-15 | 2023-09-08 | 四川三思德科技有限公司 | Intelligent community monitoring management method, system and medium based on dynamic target detection |
CN117278837A (en) * | 2023-11-16 | 2023-12-22 | 新乡天辅电子科技有限公司 | Emergency rescue-oriented imaging equipment control method |
CN117278837B (en) * | 2023-11-16 | 2024-01-26 | 新乡天辅电子科技有限公司 | Emergency rescue-oriented imaging equipment control method |
Also Published As
Publication number | Publication date |
---|---|
CN105979210B (en) | 2019-02-05 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN105979210A (en) | Pedestrian identification system based on multi-ball multi-gun camera array | |
CN112216049B (en) | Construction warning area monitoring and early warning system and method based on image recognition | |
CN105894702B (en) | A kind of intrusion detection warning system and its detection method based on multiple-camera data fusion | |
CN106650671B (en) | Face recognition method, device and system | |
CN101465033B (en) | Automatic tracking recognition system and method | |
CN103716594B (en) | Panorama splicing linkage method and device based on moving target detecting | |
CN102254169B (en) | Multi-camera-based face recognition method and multi-camera-based face recognition system | |
CN111242025B (en) | Real-time action monitoring method based on YOLO | |
CN103761514B (en) | The system and method for recognition of face is realized based on wide-angle gunlock and many ball machines | |
CN105354563B (en) | Face datection prior-warning device and implementation method are blocked in conjunction with depth and color image | |
US20090324010A1 (en) | Neural network-controlled automatic tracking and recognizing system and method | |
CN101635834A (en) | Automatic tracing identification system for artificial neural control | |
DE102008039130A1 (en) | Automatic tracing and identification system for movable object e.g. human, in building, has safety monitoring sensor connected with safety monitoring system such that tracing camera receives desired data when sensor is operated | |
KR101788225B1 (en) | Method and System for Recognition/Tracking Construction Equipment and Workers Using Construction-Site-Customized Image Processing | |
CN101814130A (en) | Iris identification device by using camera array and multimodal biometrics identification method | |
DE19955714A1 (en) | Method and apparatus for photographing / recognizing a face | |
CN103049734A (en) | Method and system for finding person in public place | |
KR20120019008A (en) | Method and devices for tracking multiple object | |
CN102930261A (en) | Face snapshot recognition method | |
CN103167270B (en) | Personnel's head image pickup method, system and server | |
KR100326203B1 (en) | Method and apparatus for face photographing and recognizing by automatic trading a skin color and motion | |
CN104966067A (en) | Image processing method and system for protecting privacy | |
CN109886129B (en) | Prompt message generation method and device, storage medium and electronic device | |
CN109241847A (en) | The Oilfield Operation District safety monitoring system of view-based access control model image | |
CN111325133A (en) | Image processing system based on artificial intelligence recognition |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20190205 Termination date: 20210606 |
|
CF01 | Termination of patent right due to non-payment of annual fee |