CN107093171A - A kind of image processing method and device, system - Google Patents
A kind of image processing method and device, system Download PDFInfo
- Publication number
- CN107093171A CN107093171A CN201610090915.5A CN201610090915A CN107093171A CN 107093171 A CN107093171 A CN 107093171A CN 201610090915 A CN201610090915 A CN 201610090915A CN 107093171 A CN107093171 A CN 107093171A
- Authority
- CN
- China
- Prior art keywords
- information
- image
- monitoring
- view
- target
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/10—Human or animal bodies, e.g. vehicle occupants or pedestrians; Body parts, e.g. hands
- G06V40/16—Human faces, e.g. facial parts, sketches or expressions
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/60—Type of objects
- G06V20/64—Three-dimensional objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10004—Still image; Photographic image
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Oral & Maxillofacial Surgery (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
- Image Processing (AREA)
- Alarm Systems (AREA)
Abstract
The embodiment of the invention discloses a kind of image processing method and device, system, wherein, methods described includes：Obtain the image information for Target monitoring area for shooting and obtaining；The monitoring object in the Target monitoring area is determined from the plane picture of image information, and the depth information in image information calculates the three-dimensional coordinate for obtaining each point for representing the monitoring object, and the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；By the overlapping 3-D view for mapping on the plane picture in described image information, obtaining the monitoring object of the threedimensional model depth map.Using the present invention, accurately and quickly a 3-D view can be obtained by way of three-dimensional modeling and two dimensional image mapping, obtained 3-D view can realize the multi-angle displaying for monitoring object, and the convenient contrast subsequently through all angles accurately completes Object identifying.
Description
Technical field
The present invention relates to technical field of image processing, more particularly to a kind of image processing method and device, system.
Background technology
Face and human body recognition technology are face feature and physical characteristic based on people, image to input or
Video carries out analysis identification, determines the position of multiple features in face and human body, size and each main device
The characteristics of image such as the positional information of official, with suitable recognizer by known to these characteristics of image and identity
Face and human body are contrasted, so as to identify the identity of each face.
With the development of image recognition technology, current face and human body recognition technology can relatively accurately be known
Do not go out who object, meet the automating of person recognition, intelligent demand, need the field of security all kinds of
It is widely applied in scape.
But, in the image processing process before carrying out contrast identification identity, also there is be difficult to overcome to lack
Fall into, many influences of such as illumination condition deficiency cause the two dimensional image feature enough from extracting,
And then reduce the accuracy rate of image object identification.
The content of the invention
The embodiment of the present invention provides a kind of image processing method and device, system, can be more accurately and effectively true
Make the 3-D view for monitoring object.
The embodiments of the invention provide a kind of image processing method, including：
Obtain the image information for Target monitoring area for shooting and obtaining；
The monitoring object in the Target monitoring area, and root are determined from the plane picture of image information
The three-dimensional coordinate for obtaining each point for representing the monitoring object, root are calculated according to the depth information in image information
The threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；
Map to the threedimensional model depth map is overlapping on the plane picture in described image information, obtain institute
State the 3-D view of monitoring object.
Correspondingly, the embodiment of the present invention additionally provides a kind of image processing apparatus, including：
Acquisition module, the obtained image information for Target monitoring area is shot for obtaining；
Determining module, for being determined from the plane picture of image information in the Target monitoring area
Monitoring object, and depth information in image information calculates each point for obtaining representing the monitoring object
Three-dimensional coordinate, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；
Processing module, for by the overlapping plane mapped in described image information of the threedimensional model depth map
On image, the 3-D view of the monitoring object is obtained.
Correspondingly, the embodiment of the present invention additionally provides a kind of image processing system, intelligent camera device and service
Device；
The intelligent camera device, for shooting the image information obtained for Target monitoring area；
The server, for obtain the intelligent camera device shoot obtain for Target monitoring area
Image information；The monitoring object in the Target monitoring area is determined from the plane picture of image information,
And the depth information in image information builds the threedimensional model depth map of the monitoring object；By described three
Dimension module depth map is overlapping to be mapped on the plane picture in described image information, obtains the monitoring object
3-D view.
The embodiment of the present invention can obtain the 3-D view on monitoring object based on the image information of shooting, its
Implementation is accurate, quick, and obtained 3-D view can realize the multi-angle displaying for monitoring object,
The convenient contrast subsequently through all angles accurately completes Object identifying.
Brief description of the drawings
Fig. 1 is a kind of schematic flow sheet of image processing method of the embodiment of the present invention；
Fig. 2 is the schematic flow sheet of another image processing method of the embodiment of the present invention；
Fig. 3 is a kind of schematic flow sheet of the method for determination Target monitoring area of the embodiment of the present invention；
Fig. 4 is a kind of structural representation of image processing system of the embodiment of the present invention；
Fig. 5 is a kind of structural representation of image processing apparatus of the embodiment of the present invention；
Fig. 6 is a kind of structural representation of server of the embodiment of the present invention.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear
Chu, it is fully described by, it is clear that described embodiment is only a part of embodiment of the invention, rather than
Whole embodiments.Based on the embodiment in the present invention, those of ordinary skill in the art are not making creation
Property work under the premise of the every other embodiment that is obtained, belong to the scope of protection of the invention.
The embodiment of the present invention can dispose intelligent camera device, and the intelligent camera device can be to some monitored area
Shot, obtain including the image information of plane picture and depth information, in order to human body therein,
The monitoring objects such as face are monitored.According to the depth information life that obtained plane picture is shot in all angles
Identify the three-dimensional coordinate of each pixel of some target monitoring object in paired monitored area, and according to obtaining
Three-dimensional coordinate build and obtain threedimensional model depth map, then threedimensional model depth map and all angles are clapped again
The plane picture for taking the photograph correspondence monitoring object in obtained image information carries out transparent and overlap processing, finally gives
The complete three-dimensional real image on the monitoring object, such as three-dimensional face or 3 D human body image pass through
The operations such as rotation, translation are it can be seen that the feature in each face of the 3-D view such as the face, human body.
Fig. 1 is referred to, is a kind of schematic flow sheet of image processing method of the embodiment of the present invention, the present invention is real
Applying the methods described of example can apply in the Object identifying field based on image, specifically can be by such as personal electricity
The equipment such as brain, server are coordinated to perform, or by the intercommunication for the multiple servers for constituting cloud platform
Perform described image processing method.Specifically, the embodiment of the present invention the described method comprises the following steps.
S101：Obtain the image information for Target monitoring area for shooting and obtaining.
Image information, 3D cameras can be obtained by configuring the 3D cameras in some particular locations
At least including a camera is used to shoot 2D images, in addition to a probe for being used to obtain depth information,
Such as infrared probe, ultrasonic probe.Specifically, 3D cameras can include RGB (red,
Green, blue, abbreviation RGB) camera and infrared ray (Infrared Radiation, abbreviation IR) camera lens,
RGB cameras are to obtain two-dimensional color figure, and IR camera lenses include two, and a camera lens is used to launch infrared light,
The infrared light that one camera lens is used to receive the transmitting runs into the light that some point back reflection of object is returned, base
The point is obtained to information such as the distances of camera lens in transmitting and the time difference received, and then obtains monitoring on some
Each three-dimensional coordinate point of object, subsequently obtained each three-dimensional coordinate point can calculate by algorithm into
And construct 3D model depth figures.Each pixel of each point of the depth map that IR camera lenses are obtained and RGB cameras
Point is generally corresponded to.
Taken image information includes plane picture and depth information, and the depth information represents the mesh
Mark the distance that the object in monitored area arrives camera, each point and plane picture in depth information each
Pixel is corresponded to substantially.
Intelligent camera device captured in real-time monitored area is triggered, the image for monitored area is obtained.The prison
Survey region and be specifically as follows the entrance areas of such as all kinds of mansions, some of place such as station, airport region.
In embodiments of the present invention, intelligent camera device works as people typically using face or human body as monitoring object
Quantity it is smaller when when being the quantity smaller (be less than some amount threshold) of monitoring object in monitored area, can
To call 3D intelligent camera devices, obtained image information (including depth is shot directly against the monitored area
Figure and plane picture) to carry out subsequent treatment.
And when the quantity of monitoring object in monitored area more (being not less than some amount threshold), can pass through
Monitored area, in the position of monitored area, is divided into multiple regions, allowed by specific algorithm according to crowd to be analyzed
The identification number in each region is less than the limitation quantity of system algorithm, then regard these regions as target monitoring area
Domain carries out the analysis identification of face or human body.Or call the directly shooting of intelligent camera device to obtain being directed to the prison
The initial image information in region is surveyed, region segmentation then is carried out to initial image information, multiple regions are obtained
Image information, then for the image information progress subsequent treatment in each region.For example in some markets, greatly
The doorway in tall building, can typically converge more people, now carry out image recognition to some people to preferably meet
And tracking etc. monitoring requirements, multiple regions can be marked off based on the picture of detection zone or shooting, with
Just preferably monitoring object is monitored.
It should be noted that the mode of above-mentioned cut zone is random, when completing face within the specific limits
Or after the analysis identification of human body, the dividing mode in the region can be released, or in another scene, according to
Demand re-starts the division in region, the recognition and tracking to each region of division is completed, so as to improve knowledge
Other success rate.
When proceeding by monitoring object analysis identification, Target monitoring area is determined, and in the S101
The image information of Target monitoring area is obtained respectively, is implemented in order to perform the present invention to each target area
The subsequent step of the image procossing of example.
S102：The monitoring object in the Target monitoring area is determined from the plane picture of image information,
And the depth information in image information calculates the three-dimensional seat for obtaining each point for representing the monitoring object
Mark, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate.
Can based on pixel value and it is preset the need for the feature of object that monitors monitoring is determined from image information
Object, can be by reflecting the size of image pixel information amplitude of variation for example, it is desired to when monitoring face
Gradient information is monitored, and gradient information is sensitive to image border, face identification method energy based on gradient information
Enough textural characteristics for effectively extracting image, by extracting the textural characteristics of image and the people of reality
Face feature features such as (including) nose, eyes, determines face.Specifically, being determined from plane picture
Going out monitoring object can be realized based on existing face recognition technology or human body recognition technology.
Intelligent camera device can be persistently tracked to some monitoring object by Object tracking and be taken pictures, and then
Shoot and obtain some on front, below side, isoazimuth degree in same monitoring object moving process
Image information, believes further according to the depth in the image information of different angles (front, side, below etc.)
Breath sets up threedimensional model depth map to monitoring object, for example, set up three-dimensional coordinate figure to people or its head, sets up
Corresponding three-dimensional (3 D) manikin or faceform.Wherein specifically, according to the monitoring object determined in plane
Location of pixels in image, can accordingly determine the monitoring object of corresponding position in depth information.
S103：Map to the threedimensional model depth map is overlapping on the plane picture in described image information,
Obtain the 3-D view of the monitoring object.
According to space reflection corresponding relation principle, can by the three-dimensional coordinate figure of foundation it is overlapping be mapped to it is same
On two dimensional image, by transparent and overlapping, one group of actual picture is formed, 3 dimensional drawing and two dimension is completed
The images match of cromogram.According to by front, side, three-dimensional below obtained by equal angular image information
Image is that an image that can show more perspective checks that user can perform rotation over the display to user
Turn, the processing such as translation to be to check the images of the different angles for the monitoring object by the 3-D view.
If for example, when what is obtained is the 3-D view on face, positive face, the side face of all angles can be checked
Deng image.
It should be noted that can be by setting an intelligent camera device to shoot some monitoring object in movement
During multiple angles image information, or be respectively provided with multiple angle positions intelligent camera device difference
The image information of corresponding angle is shot, and then obtains image of some target monitoring object in different azimuth angle
Information, to ensure three-dimensional modeling.And if when being not picked up by the image information in some orientation, then can root
According to the image information of other orientation angles, the one side for failing to obtain image information is carried out in three-dimensional modeling
Estimation, to obtain 3-D view.
Specifically, according to mapping relations principle in three dimensions, arbitrary surface can pass through camera
Mapping, is projected between a two-dimentional image, the curved surface of parallel three dimensions is overlapping be mapped to it is same
On two dimensional image, by transparent and overlapping, one group of actual picture is formed.Three dimensions is passing through camera
Project on two dimensional image, actually one group infinite many three-dimensional point is mapped to a two-dimensional points, and one
Known three-dimensional point can be tried to achieve by matrix function transformation and uniquely determine corresponding two-dimensional points.Therefore one is passed through
Serial complicated algorithm can be matched the threedimensional model depth map of acquisition with two-dimensional color figure.
Specifically, implementing for the S102 and/or S103 can be as described above by PC or service
The equipment such as device are performed, but because its processing procedure requires higher to the process resource of equipment, PC or
The equipment such as server can complete described by being interacted with cloud platform server by cloud platform server
S102 and/or S103 simultaneously returns to result to the equipment such as PC or server.
The embodiment of the present invention can obtain the 3-D view on monitoring object based on the image information of shooting, its
Implementation is accurate, quick, and obtained 3-D view can realize the multi-angle displaying for monitoring object,
The convenient contrast subsequently through all angles accurately completes Object identifying, for example, pass through plurality of human faces 3-D view
The operation such as rotation, translation, image of the face to be identified in each face (above, side etc.) can be obtained
Feature, and then conveniently there is the contrast of more characteristics of image in recognition of face, greatly improve recognition of face
Accuracy.
Fig. 2 is referred to again, is the schematic flow sheet of another image processing method of the embodiment of the present invention, this hair
The methods described of bright embodiment can be applied in the Object identifying field based on image, specifically can be by such as individual
The equipment such as people's computer, server are in communication with each other coordination to perform, or by the multiple servers for constituting cloud platform
Perform described image processing method.The embodiment of the present invention the described method comprises the following steps.
S201：Obtain the image information for Target monitoring area for shooting and obtaining.
The intelligent camera device includes 3D cameras, and taken image information includes plane picture and depth
Information is spent, the depth information represents the object in the Target monitoring area to the distance of camera, depth
Information is corresponding with each pixel of plane picture.
Multiple images information can be obtained, is filled especially by the 3D intelligent cameras being arranged at multiple calibration positions
Put to obtain image information, or image information is persistently obtained by a 3D intelligent camera device, in order to
The image information of the different visual angles for some monitoring object is obtained, for example, is being moved through for some stranger
The image information at front, side in journey, each visual angle such as below, and then convenient subsequently obtain more accurate
3-D view.Wherein, existing video tracking technology can be used for the tracking technique of the monitoring object,
, can be according to the depth in each stereo-picture after the multiple images information of same monitoring object is obtained
Information and plane picture, obtain more accurate on the monitoring by related three-dimensionalreconstruction algorithm to combine
The threedimensional model depth map and final 3-D view of object.
S202：The monitoring object in the Target monitoring area is determined from the plane picture of image information,
And the depth information in image information calculates the three-dimensional seat for obtaining each point for representing the monitoring object
Mark, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate.
S203：Map to the threedimensional model depth map is overlapping on the plane picture in described image information,
Obtain the 3-D view of the monitoring object.
S204：3-D view for the obtained monitoring object determines identity information.
The identity information act as be used for monitoring object described in unique mark.The S204 can specifically include：
The characteristics of image of the analysis identification 3-D view, according to each characteristics of image identified from including image and with
The image similar to the 3-D view is found out in the database of the identity information of image association；It will find out
The identity information of similar image association be defined as identity information for identifying the monitoring object identity.
The database of the identity information associated including image and with image can be to record in security access system
Identity database, specifically, for example in various office blocks or residential building, can exist to visitor and step on
Note system, i.e. each visitor need to provide the certificate that can be shown that its identity, such as identity card, driving license, passport
Deng, now, some devices of Security Personnel or setting can shoot or read the information of these certificates,
So as to form the database including user identity and photo, it is strange that such as identity-based card can obtain some
Photo, name, identification card number of people etc..For some places, it might even be possible to which directly configuration is on identity card
Identity database, the ID card information of the multiple users recorded, including identity are stored in the database
Photo and name, ID card No. on card, the even identity information such as address.
By in the characteristics of image and identity database of the 3-D view on such as identity card in the S204
The feature of photo is compared, and similarity is higher than into default similarity threshold and similarity highest identity
Identity information corresponding to license piece as the monitoring object identity information.
The S204 can also include：It is that the 3-D view generates identity according to preset identification number strategy
Information, the identity information is associated with the 3-D view, the identity for identifying the monitoring object.
Do not have it is above-mentioned refer to identity card database database when, or do not determined during aspect ratio pair
Go out the identity information as the monitoring object, then directly can generate an identity information for 3-D view,
The identity information may simply be a numbering, for monitoring object described in unique mark.Identification number plan
Slightly it is mainly used in numbering for some stranger configuration is unique, may be referred to the coding staff of the certificates such as identification card number
Formula is encoded, or be directly based upon the stranger entrance time, be taken for the first time position (gate,
Side door etc.), any one or more in the feature such as sex that determines of human bioequivalence encoded.
S205：The mobile message of the monitoring object is determined, the mobile message can be the monitoring pair
As the positional information or the monitoring object where not in the same time are during whole be monitored
Movement locus.The S205 is optional step.
Global position information, such as GPS location and elevation information can be defined for each intelligent camera device,
Determine that the monitoring object is relative relative to intelligent camera device further according to the depth information in image information
Position, with reference to the global position information and relative position of intelligent camera device, can obtain the monitoring object
Positional information, the positional information drawn can be a GPS and the accurate location information of height.Pass through
The calculation can obtain the position letter for the monitoring object do not determined in the same time by image information
Breath.
Or further, the S205 can also specifically include：Plane picture in image information,
Determine positional information of the monitoring object at current time.Specifically, to the plane in image information
Image is analyzed, and identifies the position feature for representing the monitoring object present position；According to position
Feature determines positional information of the monitoring object at current time, and records the monitoring determined out
The positional information of object.It can also obtain what is do not determined in the same time by image information by the processing mode
The positional information of the monitoring object.
Complete the determination of approximate location according to position feature, it is specific for example, some image information of XX moment it is flat
Face image includes the image of mansion entrance gate, then can be directly by the positional information of the monitoring object
It is described as：At XX moment entrance gates, or the XX moment includes the image on N buildings foreground, then can be direct
The positional information of the monitoring object is described as：At XX moment N buildings foreground.Certainly may be used also when determining
To be further based on completing to determine for the preset positional information of intelligent camera device, for example, described many
Individual intelligent camera device is preset in M buildings meeting room, then these intelligent camera devices of XX moment are clapped
The positional information of monitoring object in the image information taken the photograph can be described as：XX moment M building meeting rooms.
Further, when the S205 determines the movement locus of the monitoring object, it can also specifically pass through
Following steps are realized.
Trigger filming apparatus and follow the trail of the locked object in the Target monitoring area, and according to the bag of shooting
Include the locked object image information calculate the locked object in the Target monitoring area each
The positional information at moment；The movement locus of the locked object is obtained according to the positional information at each moment.
Also, when detecting the locked object not in the coverage of the filming apparatus, according to described
The movement locus of locked object, determines other filming apparatus associated with the filming apparatus；Triggering institute
State other filming apparatus and follow the trail of the locked object, and the figure of the locked object is included according to shooting
As the positional information at information calculating locked each moment of object, the positional information refers to described locked
Position of the object in the region that other described filming apparatus are monitored；Obtained according to the positional information at each moment
To the movement locus of the locked object.
Specifically, the lock object can refer to the monitored object for generating 3-D view mentioned above,
Can certainly be the specific object of other, such as some stranger, certain animal or suspicious-looking
People etc..3D cameras have lock function in itself, and some monitoring object can be locked as locked pair by it
As what it was locked is the three-dimensional coordinate point of the locked object, three pre-set in itself according to 3D cameras
Dimension coordinate (such as gps coordinate and height coordinate) and determined based on three-dimensional coordinate point described locked pair
As the relative position (relative coordinate) relative to this 3D cameras, the position of the locked object can be obtained
Confidence ceases.And then be able to can be followed the trail of when the locked object is in the monitoring range of the 3D cameras
To the movement track of the locked object, when locked object has walked out current camera head monitor scope, due to
Backstage is in advance associated multiple 3D cameras by algorithm, so true according to existing movement locus
Make after locked object goes to another 3D camera head monitor scope from a 3D camera, backstage is also
It is that can follow the trail of and find this locked object.
Further, after carrying out track following to some lock object and recording its movement locus, after convenience
The continuous track to the lock object is reduced, and the mode being had access to subsequently through inquiry, which can be simulated, restores the lock
Determine the movement locus of object (such as one stranger).
S206：The description information of the 3-D view on the monitoring object is generated, the description information includes：
The image feature information of the 3-D view, and/or associated with the 3-D view be used for identify the monitoring
The identity information of object identity.
Described image characteristic information can include features of skin colors, hair feature (shaven head, long hair etc.) and
Feature, these image feature informations and the bodies such as example bridge of the nose height, the face size of some more details
Part information as monitoring object corresponding to the 3-D view search key.
S207：By the 3-D view and its description information associated storage into preset object identification data storehouse.
S208：If receive object search request, according to the object search ask in search information from
Target image and the corresponding description information of the target image are searched out in the preset object identification data storehouse,
The searching request includes：For the characteristics of image and/or identity information scanned for.Wherein, the mesh
Logo image include plane picture in the 3-D view, the described image information that is associated with the 3-D view,
And any one or more in obtained plane picture is extracted according to the 3-D view.Wherein, it is described
It can get as needed for 3-D view correspondence that 3-D view, which extracts obtained plane picture,
Front, side, below isoplanar image.
Specifically, the intelligent video camera head can successively obtain multiple on monitoring object plane picture in order to
It is follow-up to be used when generating the 3-D view of the monitoring object, can be by these plane pictures and obtained graphics
The associated storages such as picture, description information are into preset object identification data storehouse, in order to subsequently be initiated in user
During to the search of monitoring object, the corresponding plane picture of the monitoring object can also search for.Certainly, it is described
Each plane picture on monitoring object that intelligent camera device is photographed can also be stored separately in some data
In storehouse, in order to which user is when searching for monitoring object, also plane can be searched based on the mode of picture search
Image.
S209：Return to the search result obtained according to the searching request.
Still optionally further, it will can exist in the S207 for the monitoring object corresponding to the 3-D view
The positional information and/or movement locus that each moment is determined and the 3-D view associated storage to the object
In identification database；So, the search result of the return includes：Believed according to characteristics of image and/or identity
Cease the target image searched, the description information associated with the target image that this is searched, positional information and
Any one or more in trace information.Positional information in the search result includes searching with this
Target image association each moment positional information, or associated with the target image that this is searched each when
The positional information minimum from the duration at current time in the positional information at quarter.
When requiring to look up someone, the keyword such as the above-mentioned colour of skin referred to, hair feature can be inputted,
Further it is confirmed whether it is the object required to look up in order to find target image.And further
Positional information, movement locus according to associated by 3-D view etc., it is determined that after can also determine that the needs are looked into
The motion path of the people looked for, or current position.
Fig. 3 specifically is referred to, is that a kind of flow of the method for determination Target monitoring area of the embodiment of the present invention is shown
It is intended to, the methods described of the embodiment of the present invention can correspond to above-mentioned S201, specifically, the present invention is implemented
The methods described of example comprises the following steps.
S301：Obtain and shoot obtained image for target area.
S302：Determine the quantity for the object that monitoring condition is met in the image got.
S303：If it is determined that quantity be less than default amount threshold, then pass through intelligent camera device shoot pin
To the image information of the Target monitoring area, with the image information for Target monitoring area acquired.
S304：If it is determined that quantity be not less than default amount threshold, then shot by intelligent camera device
Obtain the initial image information for the Target monitoring area；
S305：Depth map and plane picture in the initial image information is divided, multiple areas are obtained
Domain, and using the depth map and plane picture of each corresponding region as image information, with being directed to for acquiring
The image information of Target monitoring area, is easy to be handled for each image information respectively.
The whole process again to the embodiment of the present invention is described below.It is to be mounted with that a kind of intelligent 3D takes the photograph first
As head (additionally mounted can also have intelligent 2D cameras), use of all cameras with the embodiment of the present invention
It is connected in equipment such as the PCs or server for carrying out 3-D view processing, specifically can be directly connected to and match somebody with somebody
In the cloud platform for carrying out great amount of images processing put, cloud platform have recorded all information of camera, bag
Include model, position, institute monitored area and shooting angle to the monitored area etc..
When intelligent 3D cameras capture stranger's face, it can quickly shoot and upload enough image letters
Cease to cloud platform, cloud platform quickly position the position of stranger's (monitoring object) and capture face image or
Human body image, study is used for quickly detecting to face characteristic or characteristics of human body, and the strange person can be distinguished by finding out
The unique attribute of part simultaneously forms a distinctive digital code.
Meanwhile, cloud platform is automatic from different perspectives (front, side, below etc.) according to each image information
Three-dimensional coordinate figure is set up to the people of Target monitoring area or its head, corresponding 3 D human body or face mould is set up
Type.
According to space reflection corresponding relation principle, the three-dimensional coordinate figure weight that cloud platform sets up intelligent 3D cameras
It is folded to be mapped on same two dimensional image, by transparent and overlapping, one group of actual picture is formed, three are completed
The images match of stereogram and two-dimensional color figure is tieed up, the final 3-D view on monitoring object is obtained, from
And confirm the corresponding relation of object in Different Plane photo.Obtained 3-D view can be supplied to cloud platform
Fast Learning is carried out, in order to which cloud platform completes face or human bioequivalence by the effective recognizer of high accuracy,
By this more fully information analysis, the misclassification rate in identification process can be reduced and percent of pass is improved.
Positioning and range capability (depth map) by means of 3D cameras, to same strange in angular field of view
Positional information of the people in different periods image, it is recognizable to come out and associated, realize that stranger track is known
Do not track, stranger's movement locus route is reduced by algorithm, and its movement locus figure is issued into cloud platform.
Higher place is required for security protection, intelligent 3D camera arrays can be disposed, multiple cameras are carried out
Face and tracing movement track are caught without dead angle, the camera in array is subjected to efficient association with algorithm,
So as to which a complete trajectory for moving stranger is restored.
Cloud platform completes stranger's face or human bioequivalence and determined after movement locus, can combine visitor and register pipe
The personal information such as identity card, phone number is matched in reason system, specifies the identity of stranger.If can not
Associated with guest system, cloud platform will set up an independent numbering to the stranger automatically, be distinguished.Together
When, constantly improve database images data.When requiring to look up the position of people, name or volume can be passed through
Number it can quickly navigate to his position.
When there is more stranger beyond the quantity of systematic sample number analysis limitation, cloud platform can be by specific
Algorithm analysis crowd is divided into multiple regions or group, allow each region or group identification number be less than be
System limitation number, then carry out Face datection identification.Need it is clear that, this combination is random, when one
Determine to complete in scope after analysis identification, cloud platform can release combination, or in another scene, according to need
Ask and re-start combination recognition and tracking, so as to improve recognition success rate.
The embodiment of the present invention can obtain the 3-D view on monitoring object based on the image information of shooting, its
Implementation is accurate, quick, and obtained 3-D view can realize the multi-angle displaying for monitoring object,
The convenient contrast subsequently through all angles accurately completes Object identifying.And object can be monitored
Position is determined, carries out the track following of the monitoring object, the position of the monitoring object can be determined as needed
And reduce, and the search information such as identity information, feature of correlation can also be inputted to find rapidly correspondingly
The monitoring object such as people.
Image processing apparatus, server below to the embodiment of the present invention is described in detail.
Fig. 4 is referred to, is a kind of structural representation of image processing system of the embodiment of the present invention, the present invention is real
Applying the described image processing system of example includes：Intelligent camera device 401 and server 402；Intelligent camera device
401 include 3D cameras.It should be noted that representing intelligent camera device 401 by a lines in Fig. 4
It is connected with the data of server 402, intelligent camera device 401 actually can be by wired or wirelessly
It is connected with the data of server 402.Wherein intelligent camera device 401 can include multiple.
The intelligent camera device 401, for shooting the image information obtained for Target monitoring area；
The server 402, for obtain the intelligent camera device 401 shoot obtain be directed to target monitoring
The image information in region；The prison in the Target monitoring area is determined from the plane picture of image information
Object is surveyed, and depth information in image information calculates and obtains each point for representing the monitoring object
Three-dimensional coordinate, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；Will be described
Threedimensional model depth map is overlapping to be mapped on the plane picture in described image information, obtains the monitoring object
3-D view.
The intelligent camera device includes 3D cameras, and taken image information includes plane picture and depth
Information is spent, the depth information can represent the object in the Target monitoring area to the distance of camera.
Fig. 5 specifically is referred to, is a kind of structural representation of image processing apparatus of the embodiment of the present invention, this hair
The described image processing unit of bright embodiment may be provided in above-mentioned server, can specifically be arranged on individual
In the server such as computer or server, specifically, described device is included with lower module.
Acquisition module 501, the obtained image information for Target monitoring area is shot for obtaining.
Determining module 502, for being determined from the plane picture of image information in the Target monitoring area
Monitoring object, and depth information in image information calculates each that obtain representing the monitoring object
The three-dimensional coordinate of point, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate.
Processing module 503, for by the threedimensional model depth map it is overlapping map in described image information it is flat
On the image of face, the 3-D view of the monitoring object is obtained.
Still optionally further, described device can also include：First mark module 504, for analyzing identification institute
The characteristics of image of 3-D view is stated, according to each characteristics of image identified from associating including image and with image
The image similar to the 3-D view is found out in the database of identity information；By the similar figure found out
It is defined as the identity information for identifying the monitoring object identity as the identity information of association.
Still optionally further, described device can also include：Second mark module 505, for according to preset
Identification number strategy is that the 3-D view generates identity information, and the identity information is closed with the 3-D view
Connection, the identity for identifying the monitoring object.
Still optionally further, described device can also include：Position determination module 506, it is described for determining
Positional information of the monitoring object at current time；Wherein, the positional information is according to image information and/or use
Determined in the position for the intelligent camera device for shooting described image information.
Still optionally further, the position determination module 506, specifically for putting down for including in image information
Face image is analyzed, and identifying is used for the position for representing the monitoring object present position in the plane picture
Put feature；Positional information of the monitoring object at current time is determined according to position feature, and records institute
State the positional information for the monitoring object determined.
Still optionally further, the described device of the embodiment of the present invention can also include：Tracking module 507, is used for
Trigger filming apparatus and follow the trail of the locked object in the Target monitoring area, and institute is included according to shooting
The image information for stating locked object calculates described locked object each moment in the Target monitoring area
Positional information；The movement locus of the locked object is obtained according to the positional information at each moment.
Alternatively, the tracking module 507, is additionally operable to that the locked object ought be detected not in the shooting
When in the coverage of device, according to the movement locus of the locked object, determine and filled with described shoot
Put other filming apparatus of association；Triggering other filming apparatus tracking locked object, and according to
The image information including the locked object shot calculates the position letter at locked each moment of object
Breath, the positional information refers to position of the locked object in the region that other described filming apparatus are monitored
Put；The movement locus of the locked object is obtained according to the positional information at each moment.
Still optionally further, described device can also include：Associated storage module 508, for generating on institute
The description information of the 3-D view of monitoring object is stated, the description information includes：The image of the 3-D view
Characteristic information, and/or the identity information for being used to identify the monitoring object identity associated with the 3-D view；
By the 3-D view and its description information associated storage into preset object identification data storehouse.
Still optionally further, described device can also include：Module 509 is retrieved, for searching receiving object
When rope is asked, then the search information in being asked according to the object search is from the preset object identification data
3-D view and the corresponding description information of the 3-D view are searched out in storehouse, the searching request includes：With
In the characteristics of image and/or identity information that scan for；Return to the search result obtained according to the searching request.
Still optionally further, the associated storage module 508, is additionally operable to corresponding to the 3-D view
The positional information that monitoring object is determined at each moment is known with the 3-D view associated storage to the object
In other database；The search result of the return includes：Searched according to characteristics of image and/or identity information
3-D view, the description information and positional information that are associated with the 3-D view that this is searched.Wherein, it is described
Positional information in search result includes the position letter at each moment associated with the 3-D view that this is searched
Breath, or the duration in the positional information at each moment associated with the 3-D view that this is searched from current time
Minimum positional information.
Still optionally further, described device can also include：Regional processing module 510, mesh is directed to for obtaining
Mark region shoots obtained image；Determine the number for the object that monitoring condition is met in the image got
Amount；If it is determined that quantity be less than default amount threshold, then shot by intelligent camera device for described
The image information of Target monitoring area.
Still optionally further, the regional processing module 510, be additionally operable to if it is determined that quantity not less than default
Amount threshold, then pass through intelligent camera device shoot obtain for the Target monitoring area initial pictures
Information；Depth map and plane picture in the initial image information is divided, multiple regions are obtained,
And using the depth map and plane picture of each corresponding region as image information, in order to respectively for each figure
As information is handled.
It should be noted that implementing for each functional module can join in the described device of the embodiment of the present invention
Examine the specific descriptions of corresponding steps in the method embodiment corresponding to Fig. 1 to Fig. 3.
Fig. 6 is referred to again, is a kind of structural representation of server of the embodiment of the present invention, the embodiment of the present invention
The server be specifically as follows the equipment such as PC, server, specifically, the server includes：
Input interface 601, processor 602 and memory 603.The input interface 601 respectively with one or
Multiple intelligent camera devices are connected, and realize the data transfer between the processor 602 and intelligent camera device.
It is stored with the memory 603 and carries out the application program of image procossing.The input interface 601, processor
It is connected between 602 and memory 603 by bus 604.
The memory 603 can include volatile memory 603 (volatile memory), such as arbitrary access
Memory 603 (random-access memory, RAM)；Memory 603 can also include non-volatile
Memory 603 (non-volatile memory), such as flash memory 603 (flash memory)；Deposit
Reservoir 603 can also include the combination of the memory 603 of mentioned kind.
The processor 602 can be central processing unit 602 (central processing unit, CPU).
Alternatively, the memory 603 is additionally operable to storage program instruction, and the processor 602 is by calling
The application program stored in the memory 603 performs corresponding image procossing.
Specifically, the processor 602, the obtained image letter for Target monitoring area is shot for obtaining
Breath；The monitoring object in the Target monitoring area, and root are determined from the plane picture of image information
The three-dimensional coordinate for obtaining each point for representing the monitoring object, root are calculated according to the depth information in image information
The threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；By the threedimensional model depth
The overlapping 3-D view for mapping on the plane picture in described image information, obtaining the monitoring object of figure.
Still optionally further, the processor 602, is additionally operable to the characteristics of image of the analysis identification 3-D view,
Searched according to each characteristics of image identified from the database of the identity information associated including image and with image
Go out the image similar to the 3-D view；The identity information of the similar image association found out is defined as
Identity information for identifying the monitoring object identity.
Still optionally further, the processor 602, it is described three to be additionally operable to according to preset identification number strategy
Image generation identity information is tieed up, the identity information is associated with the 3-D view, for identifying the monitoring
The identity of object.
Still optionally further, the processor 602, is additionally operable to determine the monitoring object at current time
Positional information；Wherein, the positional information is according to image information and/or for shooting described image information
The position of intelligent camera device is determined.
Still optionally further, the processor 602, for determining the monitoring object at current time
During positional information, specifically for analyzing the plane picture included in image information, identify described
It is used for the position feature for representing the monitoring object present position in plane picture；Determined according to position feature
Positional information of the monitoring object at current time, and the position for the monitoring object determined described in record
Confidence ceases.
Still optionally further, the processor 602, is additionally operable to triggering filming apparatus and follows the trail of in the target monitoring
Locked object in region, and according to being calculated the image information including the locked object of shooting
The positional information at locked object each moment in the Target monitoring area；According to the position at each moment
Information obtains the movement locus of the locked object.
Still optionally further, the processor 602, is additionally operable to that the locked object ought be detected not described
When in the coverage of filming apparatus, according to the movement locus of the locked object, determine and clapped with described
Take the photograph other filming apparatus of device association；Other described filming apparatus of triggering follow the trail of the locked object, and
The position at locked each moment of object is calculated according to the image information including the locked object of shooting
Confidence ceases, and the positional information refers to the locked object in the region that other described filming apparatus are monitored
Position；The movement locus of the locked object is obtained according to the positional information at each moment.
Still optionally further, the processor 602, is additionally operable to generate the 3-D view on the monitoring object
Description information, the description information includes：The image feature information of the 3-D view, and/or with it is described
The identity information for being used to identify the monitoring object identity of 3-D view association；By the 3-D view and its
Description information associated storage is into preset object identification data storehouse.
Still optionally further, the processor 602, if being additionally operable to receive object search request, according to institute
The search information stated in object search request searches out target image from the preset object identification data storehouse
And the corresponding description information of the target image, the searching request includes：Image for scanning for is special
Levy and/or identity information；The search result obtained according to the searching request is returned, wherein, the target figure
As include the plane picture in the 3-D view, the described image information that is associated with the 3-D view, with
And any one or more in obtained plane picture is extracted according to the 3-D view.
Still optionally further, the processor 602, is additionally operable to for the monitoring pair corresponding to the 3-D view
As the positional information and/or movement locus and the 3-D view associated storage determined at each moment are described in
In object identification data storehouse；The search result of the return includes：According to characteristics of image and/or identity information
The target image searched, the description information associated with the target image that this is searched, positional information and rail
Any one or more in mark information.
Still optionally further, the positional information in the search result includes closing with the target image that this is searched
The positional information at each moment of connection, or the position at each moment associated with the target image that this is searched are believed
The positional information minimum from the duration at current time in breath.
Still optionally further, the processor 602, obtained figure is shot specifically for obtaining for target area
Picture；Determine the quantity for the object that monitoring condition is met in the image got；If it is determined that quantity
Less than default amount threshold, then the image for the Target monitoring area is shot by intelligent camera device
Information.
Still optionally further, the processor 602, be specifically additionally operable to if it is determined that quantity not less than default
Amount threshold, then shoot the initial pictures letter obtained for the Target monitoring area by intelligent camera device
Breath；Depth map and plane picture in the initial image information is divided, multiple regions are obtained, and
Using the depth map and plane picture of each corresponding region as image information, in order to be directed to each image respectively
Information is handled.
It should be noted that implementing for the processor of the embodiment of the present invention refers to Fig. 1 to Fig. 3
The specific descriptions of corresponding steps in corresponding method embodiment.
The embodiment of the present invention can obtain the 3-D view on monitoring object based on the image information of shooting, its
Implementation is accurate, quick, and obtained 3-D view can realize the multi-angle displaying for monitoring object,
The convenient contrast subsequently through all angles accurately completes Object identifying.And object can be monitored
Position is determined, carries out the track following of the monitoring object, the position of the monitoring object can be determined as needed
And reduce, and the search information such as identity information, feature of correlation can also be inputted to find rapidly correspondingly
The monitoring object such as people.
One of ordinary skill in the art will appreciate that all or part of flow in above-described embodiment method is realized,
It can be by computer program to instruct the hardware of correlation to complete, described program can be stored in a calculating
In machine read/write memory medium, the program is upon execution, it may include such as the flow of the embodiment of above-mentioned each method.
Wherein, described storage medium can for magnetic disc, CD, read-only memory (Read-Only Memory,
) or random access memory (Random Access Memory, RAM) etc. ROM.
Above disclosed is only a kind of preferred embodiment of the invention, can not limit this hair with this certainly
Bright interest field, one of ordinary skill in the art will appreciate that realizing all or part of stream of above-described embodiment
Journey, and the equivalent variations made according to the claims in the present invention, still fall within and invent covered scope.
Claims (27)
1. a kind of image processing method, it is characterised in that including：
Obtain the image information for Target monitoring area for shooting and obtaining；
The monitoring object in the Target monitoring area, and root are determined from the plane picture of image information
The three-dimensional coordinate for obtaining each point for representing the monitoring object, root are calculated according to the depth information in image information
The threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；
Map to the threedimensional model depth map is overlapping on the plane picture in described image information, obtain institute
State the 3-D view of monitoring object.
2. the method as described in claim 1, it is characterised in that also include：
The characteristics of image of the analysis identification 3-D view, according to each characteristics of image identified from including image
The image similar to the 3-D view is found out in the database of the identity information associated with image；
The identity information of the similar image association found out is defined as being used to identify the monitoring object identity
Identity information.
3. the method as described in claim 1, it is characterised in that also include：
According to preset identification number strategy be the 3-D view generate identity information, the identity information with
The 3-D view association, the identity for identifying the monitoring object.
4. the method as described in any one of claims 1 to 3, it is characterised in that also include：
Determine positional information of the monitoring object at current time；
Wherein, the positional information is taken the photograph according to image information and/or for the intelligence for shooting described image information
Determined as the position of device.
5. method as claimed in claim 4, it is characterised in that described to determine that the monitoring object is being worked as
The positional information at preceding moment, including：
The plane picture included in image information is analyzed, identifying is used for table in the plane picture
Show the position feature of the monitoring object present position；
Positional information of the monitoring object at current time is determined according to position feature, and records described true
The positional information for the monitoring object made.
6. the method as described in any one of claims 1 to 3, it is characterised in that also include：
Trigger filming apparatus and follow the trail of the locked object in the Target monitoring area, and according to the bag of shooting
Include the locked object image information calculate the locked object in the Target monitoring area each
The positional information at moment；
The movement locus of the locked object is obtained according to the positional information at each moment.
7. method as claimed in claim 6, it is characterised in that also include：
When detecting the locked object not in the coverage of the filming apparatus, according to the quilt
The movement locus of lock object, determines other filming apparatus associated with the filming apparatus；
Other described filming apparatus of triggering follow the trail of the locked object, and include described locked according to shooting
The image information for determining object calculates the positional information at locked each moment of object, and the positional information refers to
Position of the locked object in the region that other described filming apparatus are monitored；
The movement locus of the locked object is obtained according to the positional information at each moment.
8. the method as described in any one of claims 1 to 3, it is characterised in that also include：
The description information of the 3-D view on the monitoring object is generated, the description information includes：It is described
The image feature information of 3-D view, and/or associated with the 3-D view be used for identify the monitoring object
The identity information of identity；
By the 3-D view and its description information associated storage into preset object identification data storehouse.
9. method as claimed in claim 8, it is characterised in that also include：
If receiving object search request, the search information in being asked according to the object search is from described pre-
Target image and the corresponding description information of the target image are searched out in the object identification data storehouse put, it is described to search
Rope request includes：For the characteristics of image and/or identity information scanned for；
Return to the search result obtained according to the searching request；
Wherein, the described image that the target image includes the 3-D view, associated with the 3-D view
Plane picture in information and any one in obtained plane picture is extracted according to the 3-D view
Or it is a variety of.
10. method as claimed in claim 9, it is characterised in that also include：
The positional information and/or fortune that will be determined for the monitoring object corresponding to the 3-D view at each moment
Dynamic rail mark is with the 3-D view associated storage into the object identification data storehouse；
The search result of the return includes：The target figure searched according to characteristics of image and/or identity information
It is any in picture, description information, positional information and the trace information associated with the target image that this is searched
It is one or more.
11. method as claimed in claim 10, it is characterised in that the positional information in the search result
Including the positional information at each moment associated with the target image that this is searched, or the target searched with this
The positional information minimum from the duration at current time in the positional information at each moment of image association.
12. the method as described in claim 1, it is characterised in that what the acquisition shooting was obtained is directed to mesh
The image information of monitored area is marked, including：
Obtain and shoot obtained image for target area；
Determine the quantity for the object that monitoring condition is met in the image got；
If it is determined that quantity be less than default amount threshold, then shot by intelligent camera device for described
The image information of Target monitoring area.
13. method as claimed in claim 12, it is characterised in that what the acquisition shooting was obtained is directed to mesh
The image information of monitored area is marked, in addition to：
If it is determined that quantity be not less than default amount threshold, then pass through intelligent camera device shoot obtain pin
To the initial image information of the Target monitoring area；
Depth map and plane picture in the initial image information is divided, multiple regions are obtained, and
Using the depth map and plane picture of each corresponding region as image information, in order to be directed to each image respectively
Information is handled.
14. a kind of image processing apparatus, it is characterised in that including：
Acquisition module, the obtained image information for Target monitoring area is shot for obtaining；
Determining module, for being determined from the plane picture of image information in the Target monitoring area
Monitoring object, and depth information in image information calculates each point for obtaining representing the monitoring object
Three-dimensional coordinate, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；
Processing module, for by the overlapping plane mapped in described image information of the threedimensional model depth map
On image, the 3-D view of the monitoring object is obtained.
15. device as claimed in claim 14, it is characterised in that also include：
First mark module, the characteristics of image for analyzing the identification 3-D view is each according to what is identified
Characteristics of image is found out and the graphics from the database of the identity information associated including image and with image
As similar image；The identity information of the similar image association found out is defined as being used to identify the prison
Survey the identity information of object identity.
16. device as claimed in claim 14, it is characterised in that also include：
Second mark module, for being 3-D view generation identity letter according to preset identification number strategy
Breath, the identity information is associated with the 3-D view, the identity for identifying the monitoring object.
17. the device as described in any one of claim 14 to 16, it is characterised in that also include：
Position determination module, for determining positional information of the monitoring object at current time；
Wherein, the positional information is taken the photograph according to image information and/or for the intelligence for shooting described image information
Determined as the position of device.
18. device as claimed in claim 17, it is characterised in that
The position determination module, specifically for analyzing the plane picture included in image information,
Identifying is used for the position feature for representing the monitoring object present position in the plane picture；According to position
Feature determines positional information of the monitoring object at current time, and the prison determined described in record
Survey the positional information of object.
19. the device as described in any one of claim 14 to 16, it is characterised in that also include：
Tracking module, the locked object in the Target monitoring area is followed the trail of for triggering filming apparatus,
And the locked object is calculated in the target according to the image information including the locked object of shooting
The positional information at the moment of each in monitored area；Described locked pair is obtained according to the positional information at each moment
The movement locus of elephant.
20. device as claimed in claim 19, it is characterised in that
The tracking module, is additionally operable to that the locked object ought be detected not in the shooting of the filming apparatus
In the range of when, according to the movement locus of the locked object, determine its that associated with the filming apparatus
His filming apparatus；Other described filming apparatus of triggering follow the trail of the locked object, and including according to shooting
The image information of the locked object calculates the positional information at locked each moment of object, the position
Information refers to position of the locked object in the region that other described filming apparatus are monitored；According to each
The positional information at individual moment obtains the movement locus of the locked object.
21. the device as described in any one of claim 14 to 16, it is characterised in that also include：
Associated storage module, the description information for generating the 3-D view on the monitoring object is described
Description information includes：The image feature information of the 3-D view, and/or the use associated with the 3-D view
In the identity information for identifying the monitoring object identity；By the 3-D view and its description information associated storage
Into preset object identification data storehouse.
22. device as claimed in claim 21, it is characterised in that also include：
Module is retrieved, for when receiving object search request, then according in object search request
Search information searches out target image from the preset object identification data storehouse and the target image is corresponding
Description information, and return to the search result obtained according to the searching request；The searching request includes：
For the characteristics of image and/or identity information scanned for；Return to the search knot obtained according to the searching request
Really, wherein, the target image includes the 3-D view, the described image that is associated with the 3-D view
Plane picture in information and any one in obtained plane picture is extracted according to the 3-D view
Or it is a variety of.
23. device as claimed in claim 22, it is characterised in that
The associated storage module, be additionally operable to by for the monitoring object corresponding to the 3-D view each when
Carve the positional information determined and/or movement locus and the 3-D view associated storage to the Object identifying number
According in storehouse；
The search result of the return includes：The target figure searched according to characteristics of image and/or identity information
It is any in picture, description information, positional information and the trace information associated with the target image that this is searched
It is one or more.
24. device as claimed in claim 23, it is characterised in that the positional information in the search result
Including the positional information at each moment associated with the target image that this is searched, or the target searched with this
The positional information minimum from the duration at current time, Huo Zheyu in the positional information at each moment of image association
The movement locus of the target image searched association.
25. device as claimed in claim 14, it is characterised in that also include：
Regional processing module, obtained image is shot for obtaining for target area；Determine the acquisition
To image in meet monitoring condition object quantity；If it is determined that quantity be less than default amount threshold,
The image information for the Target monitoring area is then shot by intelligent camera device.
26. device as claimed in claim 25, it is characterised in that
The regional processing module, be additionally operable to if it is determined that quantity be not less than default amount threshold, then lead to
Cross intelligent camera device and shoot the initial image information obtained for the Target monitoring area；To described initial
Depth map and plane picture in image information are divided, and obtain multiple regions, and by each corresponding region
Depth map and plane picture as image information, in order to be handled respectively for each image information.
27. a kind of image processing system, it is characterised in that including：Intelligent camera device and server；
The intelligent camera device, for shooting the image information obtained for Target monitoring area；
The server, for obtain the intelligent camera device shoot obtain for Target monitoring area
Image information；The monitoring object in the Target monitoring area is determined from the plane picture of image information,
And the depth information in image information calculates the three-dimensional seat for obtaining each point for representing the monitoring object
Mark, the threedimensional model depth map of the monitoring object is built according to obtained three-dimensional coordinate；By the three-dimensional mould
The overlapping three-dimensional for mapping on the plane picture in described image information, obtaining the monitoring object of type depth map
Image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610090915.5A CN107093171B (en) | 2016-02-18 | 2016-02-18 | Image processing method, device and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610090915.5A CN107093171B (en) | 2016-02-18 | 2016-02-18 | Image processing method, device and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN107093171A true CN107093171A (en) | 2017-08-25 |
| CN107093171B CN107093171B (en) | 2021-04-30 |
Family
ID=59646145
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610090915.5A Active CN107093171B (en) | 2016-02-18 | 2016-02-18 | Image processing method, device and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107093171B (en) |
Cited By (35)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107563359A (en) * | 2017-09-29 | 2018-01-09 | 重庆市智权之路科技有限公司 | Recognition of face temperature analysis generation method is carried out for dense population |
| CN108550185A (en) * | 2018-05-31 | 2018-09-18 | Oppo广东移动通信有限公司 | Beautifying faces treating method and apparatus |
| CN108764180A (en) * | 2018-05-31 | 2018-11-06 | Oppo广东移动通信有限公司 | Face identification method, device, electronic equipment and readable storage medium storing program for executing |
| CN108876708A (en) * | 2018-05-31 | 2018-11-23 | Oppo广东移动通信有限公司 | Image processing method, device, electronic equipment and storage medium |
| CN108881888A (en) * | 2018-08-08 | 2018-11-23 | 盎锐（上海）信息科技有限公司 | Data processing equipment and method based on image collecting device |
| CN108876709A (en) * | 2018-05-31 | 2018-11-23 | Oppo广东移动通信有限公司 | Method for beautifying faces, device, electronic equipment and readable storage medium storing program for executing |
| CN108924497A (en) * | 2018-07-17 | 2018-11-30 | 浙江理工大学 | A kind of unmanned detecting system of 3D and method |
| CN109061049A (en) * | 2018-06-21 | 2018-12-21 | 河南天腾测绘科技有限公司 | A kind of gas data monitoring method of all region covering |
| CN109272576A (en) * | 2018-09-30 | 2019-01-25 | Oppo广东移动通信有限公司 | A kind of data processing method, MEC server, terminal device and device |
| CN109299510A (en) * | 2018-08-27 | 2019-02-01 | 百度在线网络技术（北京）有限公司 | Threedimensional model determines method, device and equipment |
| CN109769326A (en) * | 2017-11-09 | 2019-05-17 | 杭州海康威视数字技术股份有限公司 | One kind is followed spot method, device and equipment |
| CN109857904A (en) * | 2018-12-28 | 2019-06-07 | 北京市劳动保护科学研究所 | A kind of video frequency tracking method and device based on threedimensional model |
| CN109949449A (en) * | 2019-01-28 | 2019-06-28 | 平安科技（深圳）有限公司 | Identification visitor identity process, device, computer equipment based on recognition of face |
| CN110120060A (en) * | 2018-02-06 | 2019-08-13 | 广东虚拟现实科技有限公司 | Recognition methods, device and the recognition and tracking system of marker |
| CN110188616A (en) * | 2019-05-05 | 2019-08-30 | 盎锐（上海）信息科技有限公司 | Space modeling method and device based on 2D and 3D image |
| CN110232353A (en) * | 2019-06-12 | 2019-09-13 | 成都世纪光合作用科技有限公司 | A kind of method and apparatus obtaining scene personnel depth location |
| CN110472599A (en) * | 2019-08-20 | 2019-11-19 | 北京海益同展信息科技有限公司 | Number of objects determines method, apparatus, storage medium and electronic equipment |
| CN110728649A (en) * | 2018-06-28 | 2020-01-24 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating location information |
| CN110826357A (en) * | 2018-08-07 | 2020-02-21 | 北京市商汤科技开发有限公司 | Method, device, medium and equipment for three-dimensional detection and intelligent driving control of object |
| CN110956644A (en) * | 2018-09-27 | 2020-04-03 | 杭州海康威视数字技术股份有限公司 | Motion trail determination method and system |
| CN111178205A (en) * | 2019-12-20 | 2020-05-19 | 航天信息股份有限公司 | Method and system for identifying target object in area range |
| CN111210471A (en) * | 2018-11-22 | 2020-05-29 | 北京欣奕华科技有限公司 | Positioning method, device and system |
| CN111866447A (en) * | 2020-06-09 | 2020-10-30 | 合肥美的智能科技有限公司 | Demisting control method of storage cabinet and storage cabinet |
| CN111882656A (en) * | 2020-06-19 | 2020-11-03 | 深圳宏芯宇电子股份有限公司 | Graph processing method, equipment and storage medium based on artificial intelligence |
| CN112001947A (en) * | 2020-07-30 | 2020-11-27 | 海尔优家智能科技（北京）有限公司 | Shooting position determining method and device, storage medium and electronic device |
| CN112911205A (en) * | 2019-12-04 | 2021-06-04 | 上海图漾信息科技有限公司 | Monitoring system and method |
| CN113301242A (en) * | 2020-02-21 | 2021-08-24 | 中兴通讯股份有限公司 | Photographing method, terminal, and computer-readable storage medium |
| CN113494803A (en) * | 2020-05-29 | 2021-10-12 | 海信集团有限公司 | Intelligent refrigerator and storage and taking operation detection method for storage articles in refrigerator door |
| CN113518910A (en) * | 2019-01-17 | 2021-10-19 | 莫勒库莱特股份有限公司 | Modular system for multi-modality imaging and analysis |
| CN113596400A (en) * | 2021-07-28 | 2021-11-02 | 上海六梓科技有限公司 | Method for backtracking track |
| CN113614486A (en) * | 2019-01-17 | 2021-11-05 | 大学健康网络 | Systems, methods, and apparatus for three-dimensional imaging, measurement, and display of wound and tissue specimens |
| WO2021258282A1 (en) * | 2020-06-23 | 2021-12-30 | 深圳市大疆创新科技有限公司 | Target detection device and method, imaging apparatus, and mobile platform |
| CN114648614A (en) * | 2022-05-24 | 2022-06-21 | 四川中绳矩阵技术发展有限公司 | Three-dimensional reproduction method and system of target object |
| CN114712734A (en) * | 2022-05-05 | 2022-07-08 | 上海伽玛星科技发展有限公司 | Radiotherapy ray shooting device |
| CN114758305A (en) * | 2022-06-15 | 2022-07-15 | 成都西物信安智能系统有限公司 | Method for constructing intrusion early warning monitoring database |
Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005260753A (en) * | 2004-03-12 | 2005-09-22 | Ntt Docomo Inc | Device and method for selecting camera |
| WO2007008715A2 (en) * | 2005-07-07 | 2007-01-18 | Ingenious Targeting Laboratory, Inc. | System for 3d monitoring and analysis of motion behavior of targets |
| CN102254169A (en) * | 2011-08-23 | 2011-11-23 | 东北大学秦皇岛分校 | Multi-camera-based face recognition method and multi-camera-based face recognition system |
| CN102779359A (en) * | 2012-07-13 | 2012-11-14 | 南京大学 | Automatic ticket checking device for performing passage detection based on depth image |
| CN103260015A (en) * | 2013-06-03 | 2013-08-21 | 程志全 | Three-dimensional visual monitoring system based on RGB-Depth camera |
| US20140152836A1 (en) * | 2012-11-30 | 2014-06-05 | Stephen Jeffrey Morris | Tracking people and objects using multiple live and recorded surveillance camera video feeds |
| CN104182747A (en) * | 2013-05-28 | 2014-12-03 | 株式会社理光 | Object detection and tracking method and device based on multiple stereo cameras |
| CN104376577A (en) * | 2014-10-21 | 2015-02-25 | 南京邮电大学 | Multi-camera multi-target tracking algorithm based on particle filtering |
| CN104474710A (en) * | 2014-12-09 | 2015-04-01 | 山东大学 | System and method for tracking group users in large-scale scenes on basis of Kinect networks |
| CN104601878A (en) * | 2013-10-30 | 2015-05-06 | 索尼公司 | system and method for tracking objects |
| CN104902246A (en) * | 2015-06-17 | 2015-09-09 | 浙江大华技术股份有限公司 | Video monitoring method and device |
-
2016
- 2016-02-18 CN CN201610090915.5A patent/CN107093171B/en active Active
Patent Citations (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2005260753A (en) * | 2004-03-12 | 2005-09-22 | Ntt Docomo Inc | Device and method for selecting camera |
| WO2007008715A2 (en) * | 2005-07-07 | 2007-01-18 | Ingenious Targeting Laboratory, Inc. | System for 3d monitoring and analysis of motion behavior of targets |
| CN102254169A (en) * | 2011-08-23 | 2011-11-23 | 东北大学秦皇岛分校 | Multi-camera-based face recognition method and multi-camera-based face recognition system |
| CN102779359A (en) * | 2012-07-13 | 2012-11-14 | 南京大学 | Automatic ticket checking device for performing passage detection based on depth image |
| US20140152836A1 (en) * | 2012-11-30 | 2014-06-05 | Stephen Jeffrey Morris | Tracking people and objects using multiple live and recorded surveillance camera video feeds |
| CN104182747A (en) * | 2013-05-28 | 2014-12-03 | 株式会社理光 | Object detection and tracking method and device based on multiple stereo cameras |
| CN103260015A (en) * | 2013-06-03 | 2013-08-21 | 程志全 | Three-dimensional visual monitoring system based on RGB-Depth camera |
| CN104601878A (en) * | 2013-10-30 | 2015-05-06 | 索尼公司 | system and method for tracking objects |
| CN104376577A (en) * | 2014-10-21 | 2015-02-25 | 南京邮电大学 | Multi-camera multi-target tracking algorithm based on particle filtering |
| CN104474710A (en) * | 2014-12-09 | 2015-04-01 | 山东大学 | System and method for tracking group users in large-scale scenes on basis of Kinect networks |
| CN104902246A (en) * | 2015-06-17 | 2015-09-09 | 浙江大华技术股份有限公司 | Video monitoring method and device |
Non-Patent Citations (3)
| Title |
|---|
| KYUNGNAM KIM等: "Multi-camera Tracking and Segmentation of Occluded People on Ground Plane Using Search-Guided Particle Filtering", 《EUROPEAN CONFERENCE ON COMPUTER VISION》 * |
| 吴松: "基于监控视频的人脸识别与运动目标分割", 《中国优秀硕士学位论文全文数据库_信息科技辑》 * |
| 郭哲等: "三维到二维：人脸本征形状描述图", 《光学精密工程》 * |
Cited By (51)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107563359B (en) * | 2017-09-29 | 2018-09-11 | 重庆市智权之路科技有限公司 | Recognition of face temperature is carried out for dense population and analyzes generation method |
| CN107563359A (en) * | 2017-09-29 | 2018-01-09 | 重庆市智权之路科技有限公司 | Recognition of face temperature analysis generation method is carried out for dense population |
| CN109769326B (en) * | 2017-11-09 | 2020-09-04 | 杭州海康威视数字技术股份有限公司 | Light following method, device and equipment |
| CN109769326A (en) * | 2017-11-09 | 2019-05-17 | 杭州海康威视数字技术股份有限公司 | One kind is followed spot method, device and equipment |
| CN110120060B (en) * | 2018-02-06 | 2023-07-14 | 广东虚拟现实科技有限公司 | Identification method and device for marker and identification tracking system |
| CN110120060A (en) * | 2018-02-06 | 2019-08-13 | 广东虚拟现实科技有限公司 | Recognition methods, device and the recognition and tracking system of marker |
| CN108876709A (en) * | 2018-05-31 | 2018-11-23 | Oppo广东移动通信有限公司 | Method for beautifying faces, device, electronic equipment and readable storage medium storing program for executing |
| CN108550185A (en) * | 2018-05-31 | 2018-09-18 | Oppo广东移动通信有限公司 | Beautifying faces treating method and apparatus |
| CN108876708A (en) * | 2018-05-31 | 2018-11-23 | Oppo广东移动通信有限公司 | Image processing method, device, electronic equipment and storage medium |
| CN108764180A (en) * | 2018-05-31 | 2018-11-06 | Oppo广东移动通信有限公司 | Face identification method, device, electronic equipment and readable storage medium storing program for executing |
| CN109061049A (en) * | 2018-06-21 | 2018-12-21 | 河南天腾测绘科技有限公司 | A kind of gas data monitoring method of all region covering |
| CN110728649A (en) * | 2018-06-28 | 2020-01-24 | 北京京东尚科信息技术有限公司 | Method and apparatus for generating location information |
| CN108924497A (en) * | 2018-07-17 | 2018-11-30 | 浙江理工大学 | A kind of unmanned detecting system of 3D and method |
| CN110826357B (en) * | 2018-08-07 | 2022-07-26 | 北京市商汤科技开发有限公司 | Method, device, medium and equipment for three-dimensional detection and intelligent driving control of object |
| CN110826357A (en) * | 2018-08-07 | 2020-02-21 | 北京市商汤科技开发有限公司 | Method, device, medium and equipment for three-dimensional detection and intelligent driving control of object |
| CN108881888A (en) * | 2018-08-08 | 2018-11-23 | 盎锐（上海）信息科技有限公司 | Data processing equipment and method based on image collecting device |
| CN109299510A (en) * | 2018-08-27 | 2019-02-01 | 百度在线网络技术（北京）有限公司 | Threedimensional model determines method, device and equipment |
| CN109299510B (en) * | 2018-08-27 | 2020-06-30 | 百度在线网络技术（北京）有限公司 | Three-dimensional model determination method, device and equipment |
| CN110956644B (en) * | 2018-09-27 | 2023-10-10 | 杭州海康威视数字技术股份有限公司 | Motion trail determination method and system |
| CN110956644A (en) * | 2018-09-27 | 2020-04-03 | 杭州海康威视数字技术股份有限公司 | Motion trail determination method and system |
| CN109272576B (en) * | 2018-09-30 | 2023-03-24 | Oppo广东移动通信有限公司 | Data processing method, MEC server, terminal equipment and device |
| CN109272576A (en) * | 2018-09-30 | 2019-01-25 | Oppo广东移动通信有限公司 | A kind of data processing method, MEC server, terminal device and device |
| CN111210471B (en) * | 2018-11-22 | 2023-08-25 | 浙江欣奕华智能科技有限公司 | Positioning method, device and system |
| CN111210471A (en) * | 2018-11-22 | 2020-05-29 | 北京欣奕华科技有限公司 | Positioning method, device and system |
| CN109857904A (en) * | 2018-12-28 | 2019-06-07 | 北京市劳动保护科学研究所 | A kind of video frequency tracking method and device based on threedimensional model |
| CN109857904B (en) * | 2018-12-28 | 2022-02-18 | 北京市劳动保护科学研究所 | Video tracking method and device based on three-dimensional model |
| CN113518910A (en) * | 2019-01-17 | 2021-10-19 | 莫勒库莱特股份有限公司 | Modular system for multi-modality imaging and analysis |
| CN113614486A (en) * | 2019-01-17 | 2021-11-05 | 大学健康网络 | Systems, methods, and apparatus for three-dimensional imaging, measurement, and display of wound and tissue specimens |
| CN109949449A (en) * | 2019-01-28 | 2019-06-28 | 平安科技（深圳）有限公司 | Identification visitor identity process, device, computer equipment based on recognition of face |
| CN109949449B (en) * | 2019-01-28 | 2022-10-25 | 平安科技（深圳）有限公司 | Visitor identity identification method and device based on face identification, and computer equipment |
| CN110188616A (en) * | 2019-05-05 | 2019-08-30 | 盎锐（上海）信息科技有限公司 | Space modeling method and device based on 2D and 3D image |
| CN110188616B (en) * | 2019-05-05 | 2023-02-28 | 上海盎维信息技术有限公司 | Space modeling method and device based on 2D and 3D images |
| CN110232353A (en) * | 2019-06-12 | 2019-09-13 | 成都世纪光合作用科技有限公司 | A kind of method and apparatus obtaining scene personnel depth location |
| CN110472599B (en) * | 2019-08-20 | 2021-09-03 | 北京海益同展信息科技有限公司 | Object quantity determination method and device, storage medium and electronic equipment |
| CN110472599A (en) * | 2019-08-20 | 2019-11-19 | 北京海益同展信息科技有限公司 | Number of objects determines method, apparatus, storage medium and electronic equipment |
| CN112911205A (en) * | 2019-12-04 | 2021-06-04 | 上海图漾信息科技有限公司 | Monitoring system and method |
| CN112911205B (en) * | 2019-12-04 | 2023-08-29 | 上海图漾信息科技有限公司 | Monitoring system and method |
| CN111178205A (en) * | 2019-12-20 | 2020-05-19 | 航天信息股份有限公司 | Method and system for identifying target object in area range |
| CN113301242A (en) * | 2020-02-21 | 2021-08-24 | 中兴通讯股份有限公司 | Photographing method, terminal, and computer-readable storage medium |
| CN113494803A (en) * | 2020-05-29 | 2021-10-12 | 海信集团有限公司 | Intelligent refrigerator and storage and taking operation detection method for storage articles in refrigerator door |
| CN113494803B (en) * | 2020-05-29 | 2022-09-20 | 海信集团有限公司 | Intelligent refrigerator and storage and taking operation detection method for storage in refrigerator door |
| CN111866447B (en) * | 2020-06-09 | 2022-03-29 | 合肥美的智能科技有限公司 | Demisting control method of storage cabinet and storage cabinet |
| CN111866447A (en) * | 2020-06-09 | 2020-10-30 | 合肥美的智能科技有限公司 | Demisting control method of storage cabinet and storage cabinet |
| CN111882656A (en) * | 2020-06-19 | 2020-11-03 | 深圳宏芯宇电子股份有限公司 | Graph processing method, equipment and storage medium based on artificial intelligence |
| WO2021258282A1 (en) * | 2020-06-23 | 2021-12-30 | 深圳市大疆创新科技有限公司 | Target detection device and method, imaging apparatus, and mobile platform |
| CN112001947A (en) * | 2020-07-30 | 2020-11-27 | 海尔优家智能科技（北京）有限公司 | Shooting position determining method and device, storage medium and electronic device |
| CN113596400A (en) * | 2021-07-28 | 2021-11-02 | 上海六梓科技有限公司 | Method for backtracking track |
| CN114712734A (en) * | 2022-05-05 | 2022-07-08 | 上海伽玛星科技发展有限公司 | Radiotherapy ray shooting device |
| CN114712734B (en) * | 2022-05-05 | 2022-11-18 | 上海伽玛星科技发展有限公司 | Radiotherapy ray shooting device |
| CN114648614A (en) * | 2022-05-24 | 2022-06-21 | 四川中绳矩阵技术发展有限公司 | Three-dimensional reproduction method and system of target object |
| CN114758305A (en) * | 2022-06-15 | 2022-07-15 | 成都西物信安智能系统有限公司 | Method for constructing intrusion early warning monitoring database |
Also Published As
| Publication number | Publication date |
|---|---|
| CN107093171B (en) | 2021-04-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107093171A (en) | A kind of image processing method and device, system | |
| CN111126304B (en) | Augmented reality navigation method based on indoor natural scene image deep learning | |
| US11252329B1 (en) | Automated determination of image acquisition locations in building interiors using multiple data capture devices | |
| US10282856B2 (en) | Image registration with device data | |
| CN107247834B (en) | A kind of three dimensional environmental model reconstructing method, equipment and system based on image recognition | |
| US11003956B2 (en) | System and method for training a neural network for visual localization based upon learning objects-of-interest dense match regression | |
| US10043097B2 (en) | Image abstraction system | |
| Hu et al. | Approaches to large-scale urban modeling | |
| CN104021538B (en) | Object positioning method and device | |
| CN110443898A (en) | A kind of AR intelligent terminal target identification system and method based on deep learning | |
| CN103703758A (en) | Mobile augmented reality system | |
| CN108734194A (en) | A kind of human joint points recognition methods based on single depth map of Virtual reality | |
| CN106033601A (en) | Method and apparatus for detecting abnormal situation | |
| CN110334701A (en) | Collecting method based on deep learning and multi-vision visual under the twin environment of number | |
| Yu et al. | Intelligent visual-IoT-enabled real-time 3D visualization for autonomous crowd management | |
| WO2020015501A1 (en) | Map construction method, apparatus, storage medium and electronic device | |
| Ali et al. | Single image Façade segmentation and computational rephotography of House images using deep learning | |
| Li et al. | 3D map system for tree monitoring in hong kong using google street view imagery and deep learning | |
| Xia et al. | Precise indoor localization with 3D facility scan data | |
| US20220398804A1 (en) | System for generation of three dimensional scans and models | |
| Pozo et al. | 3D reconstruction technologies for using in dangerous environments with lack of light: a comparative analysis | |
| Khan et al. | A review of benchmark datasets and training loss functions in neural depth estimation | |
| CN109702747A (en) | A kind of robot dog system and its implementation | |
| CN109492513A (en) | The face space De-weight method of light field monitoring | |
| Radanovic et al. | Virtual Element Retrieval in Mixed Reality |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |