CN108764215A - Target search method for tracing, system, service centre and terminal based on video - Google Patents
Target search method for tracing, system, service centre and terminal based on video Download PDFInfo
- Publication number
- CN108764215A CN108764215A CN201810643119.9A CN201810643119A CN108764215A CN 108764215 A CN108764215 A CN 108764215A CN 201810643119 A CN201810643119 A CN 201810643119A CN 108764215 A CN108764215 A CN 108764215A
- Authority
- CN
- China
- Prior art keywords
- target
- video data
- search
- service
- video
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Multimedia (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a kind of target search method for tracing based on video, including：Service centre receives the searching request and video data that terminal is sent；Wherein, searching request includes：Search for target and service request；Service request includes：At least one of target search, target tracking, Address Recognition, circuit recommendation, real-time pictures；Video data is input to deep learning network trained in advance and carries out target identification, obtains target identification result；Service processing is carried out to target identification result according to service request, obtains corresponding information on services.This method utilizes artificial intelligence theory, it is realized in video data to specifying the automatic search of target to track by deep learning method, corresponding service processing is carried out to target identification result according to the service request of user, you can obtain the information on services for meeting user demand.The invention also discloses a kind of service centre, a kind of terminal and a kind of target search tracing systems based on video.
Description
Technical field
The present invention relates to field of image recognition, more particularly to a kind of target search method for tracing, a kind of clothes based on video
Business center, a kind of terminal and a kind of target search tracing system based on video.
Background technology
Currently, the search for needing that specified target is carried out by video can be usually encountered in life, for example police pass through
Monitor video carries out search, the search that related person segment is carried out to movie and television play and the tourist of suspect's trace by looking into
Related street image is ask to being presently in the inquiry etc. of position.And these search works go to solve by manual labor entirely substantially, no
Only staff brings a large amount of labour burden and error rate can also increase and gradually rise at any time.
Therefore, how to realize in video data and the search for specifying target is tracked automatically, meet the needs of users, be this
Field technology personnel's technical issues that need to address.
Invention content
The object of the present invention is to provide a kind of target search method for tracing based on video, this method utilize artificial intelligence reason
By being realized in video data by deep learning method to specifying the automatic search of target to track, wanted according to the service of user
It asks and corresponding service processing is carried out to target identification result, you can obtain the information on services for meeting user demand；The present invention's is another
One purpose is to provide a kind of service centre, a kind of terminal and the target search tracing system based on video.
In order to solve the above technical problems, the present invention provides a kind of target search method for tracing based on video, including：
Service centre receives the searching request and video data that terminal is sent；Wherein, described search request include：Search
Target and service request；The service request includes：Target search, Address Recognition, circuit recommendation, is drawn target tracking in real time
At least one of face；
The video data is input to deep learning network trained in advance and carries out target identification, obtains target identification knot
Fruit；
Service processing is carried out to the target identification result according to the service request, obtains corresponding information on services.
Preferably, the video data is input to deep learning network progress target identification trained in advance includes：
The deep learning network trained in advance is screened according to the service request, obtains the service request
Corresponding first network model；
The video data is input in the first network model and carries out target identification, obtains target identification result；
Then carrying out service processing to the target identification result according to the service request is specially：Pass through first net
Network model carries out corresponding service processing to the target identification result, obtains information on services.
Preferably, the video data is input to deep learning network progress target identification trained in advance includes：
Type division is carried out to described search target according to the target type divided in advance, obtains search target type；
The video data is input in the corresponding deep learning network of described search target type, target identification is obtained
As a result.
Preferably, the described video data is input to before deep learning network trained in advance carries out target identification is gone back
Including：
To the predetermined pretreatment operation of the video data.
Preferably, described to include to the predetermined pretreatment operation of the video data：
When the format difference of video file in the video data, the video data is carried out at unified format conversion
Reason；
Corresponding video processing is carried out to the video data according to the service request；
Tag definition is carried out to the video file according to the content of the video data.
The present invention provides a kind of service centre, including：
First receiving unit, searching request and video data for receiving terminal transmission；Wherein, described search is asked
Including：Search for target and service request；The service request includes：Target search, target tracking, Address Recognition, circuit push away
It recommends, at least one of real-time pictures；
Object-recognition unit carries out target knowledge for the video data to be input to deep learning network trained in advance
Not, target identification result is obtained；
Service unit is obtained for carrying out service processing to the target identification result according to the service request
Corresponding information on services.
The present invention provides a kind of target search method for tracing based on video, including：
Terminal receives the searching request that user sends；Wherein, described search request include：Search for target；
It is asked to collect video data according to described search；
Described search request and the video data are sent to service centre, so that the service centre passes through depth
Learning network asks to carry out target identification to the video data according to described search, generates information on services.
Preferably, the target search method for tracing based on video further includes：
Terminal asks to send the inquiry instruction of video data to service centre according to described search, so as to the service centre
Video data in database is searched according to the inquiry instruction.
The present invention provides a kind of terminal, including：
Request reception unit, the searching request for receiving user's transmission；Wherein, described search request include：Search for mesh
Mark；
Data collection unit collects video data for being asked according to described search；
First transmission unit, for described search request and the video data to be sent to service centre, with toilet
It states service centre and target identification is carried out to the video data according to described search request by deep learning network, generate service
Information.
The present invention provides a kind of target search tracing system based on video, including：
Terminal, the searching request for receiving user's transmission；Wherein, described search request include：Search for target；According to institute
It states searching request and collects video data；Described search request and the video data are sent to service centre, so as to described
Service centre asks to carry out target identification to the video data according to described search by deep learning network, generates service letter
Breath；
Service centre, searching request and video data for receiving terminal transmission；Wherein, described search request bag
It includes：Search for target and service request；The service request includes：Target search, target tracking, Address Recognition, circuit recommendation,
At least one of real-time pictures；The video data is input to deep learning network trained in advance and carries out target identification, is obtained
To target identification result；Service processing is carried out to the target identification result according to the service request, obtains corresponding service
Information.
Target search method for tracing provided by the present invention based on video, service centre receive the search that terminal is sent and ask
It asks and video data, searching request includes：Search for target and service request, wherein video data is terminal according to reception
User search request collect and obtain automatically, video data, which is input to deep learning network trained in advance, carries out target knowledge
Not, target identification is obtained as a result, may be implemented pair by using artificial intelligence theory, deep learning method, Video Analysis Technology
The automatic search tracking of specified target；Service processing is carried out to target identification result according to service request, can be obtained corresponding
Information on services, the handling result for the service request that information on services, that is, user specifies, for example handled when being operated for Address Recognition
To information on services be address information, the information on services obtained when being operated for target tracking is the line information of target
Deng obtaining information on services according to recognition result, you can automatic to realize the satisfaction for proposing demand for services to user.
The present invention also provides a kind of service centre, a kind of terminal and target search tracing systems based on video, have
Above-mentioned advantageous effect, details are not described herein.
Description of the drawings
In order to more clearly explain the embodiment of the invention or the technical proposal in the existing technology, to embodiment or will show below
There is attached drawing needed in technology description to be briefly described, it should be apparent that, the accompanying drawings in the following description is only this
The embodiment of invention for those of ordinary skill in the art without creative efforts, can also basis
The attached drawing of offer obtains other attached drawings.
Fig. 1 is the signaling diagram of the target search method for tracing provided in an embodiment of the present invention based on video；
Fig. 2 is the structure diagram of service centre provided in an embodiment of the present invention；
Fig. 3 is the structure diagram of terminal provided in an embodiment of the present invention；
Fig. 4 is the structure diagram of the target search tracing system provided in an embodiment of the present invention based on video.
Specific implementation mode
Core of the invention is to provide a kind of target search method for tracing based on video, and this method is managed using artificial intelligence
By being realized in video data by deep learning method to specifying the automatic search of target to track, wanted according to the service of user
It asks and corresponding service processing is carried out to target identification result, you can obtain the information on services for meeting user demand；The present invention's is another
One core is to provide a kind of service centre, a kind of terminal and the target search tracing system based on video.
In order to make the object, technical scheme and advantages of the embodiment of the invention clearer, below in conjunction with the embodiment of the present invention
In attached drawing, technical scheme in the embodiment of the invention is clearly and completely described, it is clear that described embodiment is
A part of the embodiment of the present invention, instead of all the embodiments.Based on the embodiments of the present invention, those of ordinary skill in the art
The every other embodiment obtained without making creative work, shall fall within the protection scope of the present invention.
Referring to FIG. 1, Fig. 1 is the signaling diagram of the target search method for tracing provided in this embodiment based on video；The party
Method may include：
Step S111, terminal receives the searching request that user sends.
Wherein, searching request includes：Search for target and service request.Search target refers to the target that user specifies, and services
It is required that referring to user needs the service executed to search target, for example carries out target search, target tracking, Address Recognition, circuit and push away
Recommend, real-time pictures search etc..
Search target can be the search target type instruction specified, for example searches for all people's object, searches for red pair
Shoulder packet etc.；Or specified image, such as the photo with label or one section of video etc., can be specifically shooting
One scape of street corner either some scene in designated data or someone etc..Wherein, when it is video to search for target, video
Include image information and acoustic information, can be identified in terms of two, improves the probability of success.
The type for searching for target can be personage's (foreign trade dress ornament, build profile or facial appearance etc.), the vehicles (vapour
Vehicle, motorcycle, electric vehicle, custom vehicle, ship etc.), building facilities (office building, residential building, commerce and trade street, school, fire-fighting, public lavatory,
Sign label etc.), the specific objectives such as natural land (mountains and rivers, river, lake, ocean etc.), at this to searching for the type of target not
It limits.
Step S112, terminal collects video data according to searching request.
Terminal receives and collects corresponding video data according to searching request after the searching request of user.Video data can be with
For：Real-time monitoring data, common monitoring image, individual video file etc., do not limit the source of video data at this.
The collection of video data can carry out automated intelligent collection by network, can also be sent and be searched by user
The video data etc. of rope does not limit the collection mode of video data at this.For example police carry out target to personnel A and chase after
Track works, and due to may relate to private data, can not inquire, can be inputted by police in some period on network
The monitor video of the monitoring device acquisition in somewhere is as video data；For example user B shoots streetscape near one and is used as search mesh
Mark carries out Address Recognition, and video data at this time can be the local streetscape etc. obtained by network inquiry.
Since service centre and terminal can be stored and backed up to customer data data, source video sequence can be also
Data on file including history.Due to ensure client data secure, reducing the reasons such as data search workload, in service centre
The video data in need scanned for may be stored, to reduce data search workload, reduces idle work, it is preferable that terminal
When receiving searching request, the inquiry instruction of video data is sent to service centre according to searching request, so as to service centre
Video data in database is searched according to inquiry instruction.When video data is stored in service centre, terminal no longer needs to
Video is transferred to after terminal and is retransmited to service centre, directly transmitting instruction makes service centre directly be obtained from its database
?.
In addition, address information can be directly acquired to the mobile phone of video data, video can also be directly collected.For fixation
Video collect equipment (such as location determination, known camera) address information can be obtained.For being similar to vehicle driving
Whether recorder or the video collection procedure of other dollying head classes open privacy depending on user, with allowing backstage mobile phone
Location information, but the address information can complete the confirmation of address by above-mentioned Address Recognition.
Step S113, searching request and video data are sent to service centre.
Searching request and video data are sent to service centre by terminal, so that service centre passes through deep learning network
Target identification is carried out to video data according to searching request, generates information on services.
Step S121, service centre receives the searching request and video data that terminal is sent.
Wherein, searching request can specifically include：Search for target and service request.Service request can specifically include：
At least one of target search, target tracking, Address Recognition, circuit recommendation, real-time pictures.Service centre can be provided including mesh
A variety of value-added services such as search, target tracking, Address Recognition, circuit recommendation, real-time pictures are marked, herein only with above-mentioned function services
For, service request can be according to user demand sets itself.
Step S122, video data is input to deep learning network trained in advance and carries out target identification, obtain target
Recognition result.
Deep learning frame may include TensorFlow, Caffe, Keras, CNTK, Torch7, MXNet, Leaf,
Theano, DeepLearning4, Lasagne, Neon etc., deep learning network model may include R-CNN, SPP-net,
Fast R-CNN, Faster R-CNN, LSTM and Yolo etc., deep learning network voluntarily selecting module and component can carry out net
The structure of network model, does not limit herein.
The process being trained to the deep learning network of structure does not limit, and deep learning network after training is to mesh
When target recognition capability reaches preset standard, you can carry out actual target search as trained network model and chase after
Track.Certainly, it can also be trained again, directly if when target identification result precision is relatively low in realistic objective search process
Reach user demand to recognition capability.
Specifically, training process can be：Terminal is by the video data data transmission of collection to service centre.Service centre
The data information of collection is pre-processed.Service centre's configuration specified deep learning frame and network model, and to data
Data is stored and is trained.Training result and historical summary are carried out integrated treatment by service centre, and by final processing and
Training result (network model parameter) is stored in service centre.Service centre is responsible for carrying out safety to the video data that terminal is collected
It stores and shared, ensures that public safety and individual privacy are not revealed by video data data.Only it is with above-mentioned training process herein
Example is introduced, and other training process to network model can refer to above-mentioned introduction.
The quantity of deep learning network does not limit, and corresponding with service request can set, for example, it can be set to 5 depth
Learning network is respectively used to carry out target search, target tracking, Address Recognition, circuit recommendation, the search of real-time pictures；It can also
Only after the search of one deep learning network progress target of training, the result of target search is carried out according to service request corresponding
Service processing, the i.e. executive agent of step S203 can be deep learning network, or other to execute data service
The device of processing, does not limit herein.
Preferably, the integration of service request is fast implemented to realize, promotes accuracy rate, may be used and wanted according to service
The corresponding deep learning network of structure is asked to be trained, then video data is input to deep learning network trained in advance carries out
Target identification may include：
Step 1：Deep learning network model trained in advance is screened according to service request, obtains service request
Corresponding first network model；
Step 2：Video data is input in first network model and carries out target identification, obtains target identification result；
Step 3：Then carrying out service processing to target identification result according to service request is specially：First network model pair
Target identification result carries out corresponding service processing, obtains information on services.
In addition, it is necessary to which the target searched for and tracked is of all kinds, different types of target search is that service centre brings pole
Big challenge, therefore the target type that deep learning network can also identify as needed is set, and is completed according to target type
The target search of customization and tracking.Such as when the search target type divided in advance be personage, the vehicles, building facilities with
And when natural land, person recognition network can be respectively trained, the vehicles identify network, building facilities identify network and oneself
Right landscape identification network is respectively used to identification the above-mentioned type.It can be with according to the quantitative structure of target type set depth learning network
Greatly promote the recognition accuracy to different type target, then by video data be input in advance trained deep learning network into
The step of row target identification, is specifically as follows：
1) Type division, is carried out to search target according to the target type divided in advance, obtains search target type；
2), video data is input in the corresponding deep learning network of search target type, obtains target identification result.
It is only introduced by taking above-mentioned several situations as an example herein, quantity, type in the present embodiment to deep learning network
And construction does not limit, it can be with sets itself.
Step S123, service processing is carried out to target identification result according to service request, obtains corresponding information on services.
Step S123 obtained target identification as a result, realize in video data to search for target accurate positionin,
For the service request for realizing different, need to carry out service processing to recognition result.
For example, when service request is specially target search, after search target is positioned in video data, positioning is believed
The presenting and displaying video asset information that breath and target occur is as information on services.When service request is target tracking, by all
After history and real-time monitored picture carry out target identification using deep learning video processing model, the position of label target appearance
Information, the circuit that target is generated according to the location information of label are tracked.When service request is Address Recognition, pass through the figure of shooting
Picture, such as video image confirm the specific of destination address respectively by the image, word or speech identifying function of deep learning
Label information generates position language description, including longitude and latitude, standard address etc..It, can when service request is circuit recommendation
With the navigation routine information recommended by searching for navigation information output on the basis of Address Recognition.When service request is real-time pictures
When search, the data processing that can be directed in real time data progress deep learning, to search for given target.With only with above-mentioned
For situation, the process handled target identification result according to other service requests can refer to above-mentioned introduction.
Corresponding information on services is can be obtained after carrying out service processing corresponding with preset service request.
Based on the above-mentioned technical proposal, the target search method for tracing based on video that the present embodiment is provided, service centre
Receive terminal send searching request and video data, searching request include：Search for target and service request, wherein regard
Frequency data is that according to the user search request of reception, collection obtains terminal automatically, and video data is input to depth trained in advance
It spends learning network and carries out target identification, obtain target identification as a result, by using artificial intelligence theory, deep learning method, regarding
The automatic search tracking to specifying target may be implemented in frequency analysis technology；Target identification result is serviced according to service request
Processing, can obtain corresponding information on services, the handling result for the service request that information on services, that is, user specifies, such as when for ground
The information on services handled when the identification operation of location is address information, and the information on services obtained when being operated for target tracking is i.e.
For the line information etc. of target, information on services is obtained according to recognition result, you can automatic realize proposes demand for services to user
Meet.
Based on above-described embodiment, phenomena such as there is source disunity, format disunities due to searching for obtained video, it is
The uniform characteristics extraction for ensureing deep learning network, improves recognition accuracy, it is preferable that need to collect before being identified
To video data pre-processed.If the deep learning network of training has no requirement for video file, can also
Directly the original video files of collection are directly inputted into network.It is carried out for being pre-processed before being identified herein
It introduces.
The video data for carrying out target identification in service centre by deep learning network is to be provided by pretreated video
Material, wherein preprocessing process can be completed by service centre, that is, be located in advance to the video data of collection after collecting video data
Reason, service centre is sent to by pretreated video data；It can also be pre-processed by service centre, i.e., service centre connects
The video data received is the original video data collected, and service centre is located in advance first after receiving these video datas
Reason, carries out relevant data processing work again after pretreatment.The executive agent of preprocessing process can be service centre,
It can be terminal, can be decided in its sole discretion by business model.
Preprocessing process is introduced so that the heart in service pre-processes video data as an example herein.At this to tool
The pretreatment operation of body is not construed as limiting, can sets itself as needed.
Due in video data not only can include common monitoring video, organization's monitor video, personal monitoring and control video,
Further include movie and television play video, individual recording video and satellite data video etc..Search faces first with tracking target in video
The method for processing video frequency of the problem of various video source, separate sources may be different, and format is cut and transfer process is very necessary,
So that the target video file of search and tracking obtains unification；Different service requests may not for video file requirement yet
Require the video data of search that may be necessary for 3D pictures etc.；In addition, showing mercy in a large amount of
The unknown file of condition can expend vast resources in each identification process, need at this time carry out video content tag definition and
Classifying content.
To ensure that identification process is completed rapidly and accurately, it is preferable that specifically include unified format herein with pretreatment operation
Conversion process according to service request be introduced for corresponding video processing and tag definition process.
Then the step of pretreatment behaviour predetermined to video data is specifically as follows：
One, when the format difference of video file in video data, unified format conversion processing is carried out to video data；
Two, corresponding video processing is carried out to video data according to service request；
Three, tag definition is carried out to video file according to the content of video data.
Different pre-operation processing operations can freely arrange in pairs or groups, and details are not described herein for other situations.
Referring to FIG. 2, Fig. 2 is the structure diagram of service centre provided in an embodiment of the present invention；May include：First receives
Unit 200, object-recognition unit 210 and service unit 220.Service centre provided in this embodiment can be with above-mentioned base
It is mutually compareed in the target search method for tracing of video.
Wherein, the first receiving unit 200 is mainly used for receiving the searching request and video data that terminal is sent；Wherein,
Searching request includes：Search for target and service request；Service request includes：Target search, target tracking, Address Recognition, line
At least one of road recommendation, real-time pictures.
Object-recognition unit 210 is mainly used for for video data being input to deep learning network progress target trained in advance
Identification, obtains target identification result.
Service unit 220 is mainly used for carrying out service processing to target identification result according to service request, obtains pair
The information on services answered.
Wherein it is preferred to which object-recognition unit 210 may further include：
Network screens subelement, for being screened to deep learning network trained in advance according to service request, obtains
The corresponding first network model of service request；
First identification subelement, carries out target identification for video data to be input in first network model, obtains mesh
Mark recognition result；
Then service unit 220 specifically can be used for：Target identification result is corresponded to by first network model
Service processing, obtain information on services.
Wherein it is preferred to which object-recognition unit 210 may further include：
Type division subelement is obtained for carrying out Type division to search target according to the target type divided in advance
Search for target type；
Second identification subelement, for video data to be input in the corresponding deep learning network of search target type,
Obtain target identification result.
Wherein it is preferred to which service centre may further include pretreatment unit, the input terminal of pretreatment unit and first
The output end of receiving unit connects, and the output end of pretreatment unit and the input terminal of object-recognition unit connect, for video
The predetermined pretreatment operation of data.
Preferably, pretreatment unit can specifically include：
Format analysis processing subelement, for when the format difference of video file in video data, uniting to video data
One format conversion processing；
Service processing subelement, for carrying out corresponding video processing to video data according to service request；
Tag definition subelement, for carrying out tag definition to video file according to the content of video data.
Service centre provided in this embodiment includes calculation processing resource (model training and target reasoning), net by providing
The services such as the road communication resource (video input and result output) and storage resource (model parameter and video data), in video data
It is middle to realize to specifying the automatic search of target to track, it can meet the needs of users.
Referring to FIG. 3, Fig. 3 is the structure diagram of terminal provided in this embodiment, including：Request reception unit 300, data
Collector unit 310 and the first transmission unit 320.Terminal provided in this embodiment can refer to the above-mentioned target search based on video
The introduction of method for tracing.
Wherein, request reception unit 300 is mainly used for receiving the searching request that user sends；Wherein, searching request includes：
Search for target；
Data collection unit 310 is mainly used for collecting video data according to searching request；
First transmission unit 320 is mainly used for searching request and video data being sent to service centre, to service
Target identification is carried out to video data according to searching request centrally through deep learning network, generates information on services.
Wherein it is preferred to which terminal provided in this embodiment may further include query unit, for according to searching request
To service centre send video data inquiry instruction, so as to service centre according to inquiry instruction to video data in database into
Row is searched.
The information exchange function with user may be implemented in terminal provided in this embodiment, and the service by receiving user is wanted
It asks, collects presenting and displaying video asset information, coordinating with service centre can realize in video data to specifying the automatic search of target to chase after
Track is met the needs of users.
Referring to FIG. 4, Fig. 4 is the structure diagram of the target search tracing system provided in this embodiment based on video；It should
System may include：Terminal 400 and service centre 410.
Wherein, terminal 400 primarily can be used for receiving the searching request that user sends；Wherein, searching request includes：Search
Target；Video data is collected according to searching request；Searching request and video data are sent to service centre, in servicing
The heart carries out target identification according to searching request by deep learning network to video data, generates information on services.
Terminal 400 includes not only common monitoring video, organization's monitor video, individual to the collection of video data data
Monitor video can also include movie and television play video, individual recording video and satellite data video etc..Terminal 400 can with it is public
The other-ends video capture device connections such as video monitoring, personal picture pick-up device, to realize the collection to video data.
In addition, terminal 400 can also complete the pretreatment to video data.Specifically, terminal video data data is pre-
Processing includes not only that the format of video data cuts and converts, and can also include the tag definition and classifying content of video content
Deng.
Service centre 410 primarily can be used for receiving the searching request and video data that terminal is sent；Wherein, search is asked
Ask including：Search for target and service request；Service request includes：Target search, target tracking, Address Recognition, circuit recommendation,
At least one of real-time pictures；Video data is input to deep learning network trained in advance and carries out target identification, obtains mesh
Mark recognition result；Service processing is carried out to target identification result according to service request, obtains corresponding information on services.
Service centre 410 further includes database, can be used for the storage to data information, including video data storage and
Backup further includes storage and backup and the storage of customer data data and the backup of training pattern parameter, passes through safety, concentrates
Data center protect privacy, to achieve the purpose that enhance security protection, firm society, protection individual privacy.
After obtaining information on services by service centre 410, to ensure acquisition of the user to information on services, need to believe service
Breath is exported, and is not limited output device at this,
Output device specifically can output information to client's equipped with voice output or word output etc.
Component.Intelligent terminal can be the smart machines such as mobile phone, tablet, or equipped with the terminal 400 of output block.
Target search tracing system based on video can refer to the introduction of above-mentioned terminal and service centre, no longer superfluous herein
It states.
Target search tracing system provided in this embodiment based on video is received user instructions by terminal, collects video
Data is realized in video data to specifying the automatic search of target to track, according to the service request of user by service centre
Corresponding service processing is carried out to target identification result, you can obtain the information on services for meeting user demand, the two cooperation can be with
It realizes the automatic target search based on video, meets the various demands of user.
Each embodiment is described by the way of progressive in specification, the highlights of each of the examples are with other realities
Apply the difference of example, just to refer each other for identical similar portion between each embodiment.For device disclosed in embodiment
Speech, since it is corresponded to the methods disclosed in the examples, so description is fairly simple, related place is referring to method part illustration
?.
Professional further appreciates that, unit described in conjunction with the examples disclosed in the embodiments of the present disclosure
And algorithm steps, can be realized with electronic hardware, computer software, or a combination of the two, in order to clearly demonstrate hardware and
The interchangeability of software generally describes each exemplary composition and step according to function in the above description.These
Function is implemented in hardware or software actually, depends on the specific application and design constraint of technical solution.Profession
Technical staff can use different methods to achieve the described function each specific application, but this realization is not answered
Think beyond the scope of this invention.
The step of method described in conjunction with the examples disclosed in this document or algorithm, can directly be held with hardware, processor
The combination of capable software module or the two is implemented.Software module can be placed in random access memory (RAM), memory, read-only deposit
Reservoir (ROM), electrically programmable ROM, electrically erasable ROM, register, hard disk, moveable magnetic disc, CD-ROM or technology
In any other form of storage medium well known in field.
Above the target search method for tracing to provided by the present invention based on video, terminal, service centre, be based on video
Target search tracing system be described in detail.Specific case used herein is to the principle of the present invention and embodiment
It is expounded, the explanation of above example is only intended to facilitate the understanding of the method and its core concept of the invention.It should be pointed out that
For those skilled in the art, without departing from the principle of the present invention, can also to the present invention into
Row some improvements and modifications, these improvement and modification are also fallen within the protection scope of the claims of the present invention.
Claims (10)
1. a kind of target search method for tracing based on video, which is characterized in that including：
Service centre receives the searching request and video data that terminal is sent；Wherein, described search request include：Search for target
And service request；The service request includes：In target search, target tracking, Address Recognition, circuit recommendation, real-time pictures
It is at least one；
The video data is input to deep learning network trained in advance and carries out target identification, obtains target identification result；
Service processing is carried out to the target identification result according to the service request, obtains corresponding information on services.
2. the target search method for tracing based on video as described in claim 1, which is characterized in that the video data is defeated
Enter to deep learning network progress target identification trained in advance and includes：
The deep learning network trained in advance is screened according to the service request, the service request is obtained and corresponds to
First network model；
The video data is input in the first network model and carries out target identification, obtains target identification result；
Then carrying out service processing to the target identification result according to the service request is specially：Pass through the first network mould
Type carries out corresponding service processing to the target identification result, obtains information on services.
3. the target search method for tracing based on video as described in claim 1, which is characterized in that the video data is defeated
Enter to deep learning network progress target identification trained in advance and includes：
Type division is carried out to described search target according to the target type divided in advance, obtains search target type；
The video data is input in the corresponding deep learning network of described search target type, target identification knot is obtained
Fruit.
4. the target search method for tracing based on video as described in claim 1, which is characterized in that described to provide the video
Material is input to before deep learning network trained in advance carries out target identification：
To the predetermined pretreatment operation of the video data.
5. the target search method for tracing based on video as claimed in claim 4, which is characterized in that described to be provided to the video
Expect that predetermined pretreatment operation includes：
When the format difference of video file in the video data, unified format conversion processing is carried out to the video data；
Corresponding video processing is carried out to the video data according to the service request；
Tag definition is carried out to the video file according to the content of the video data.
6. a kind of service centre, which is characterized in that including：
First receiving unit, searching request and video data for receiving terminal transmission；Wherein, described search request bag
It includes：Search for target and service request；The service request includes：Target search, target tracking, Address Recognition, circuit recommendation,
At least one of real-time pictures；
Object-recognition unit carries out target identification for the video data to be input to deep learning network trained in advance,
Obtain target identification result；
Service unit is corresponded to for carrying out service processing to the target identification result according to the service request
Information on services.
7. a kind of target search method for tracing based on video, which is characterized in that including：
Terminal receives the searching request that user sends；Wherein, described search request include：Search for target；
It is asked to collect video data according to described search；
Described search request and the video data are sent to service centre, so that the service centre passes through deep learning
Network asks to carry out target identification to the video data according to described search, generates information on services.
8. the target search method for tracing based on video as claimed in claim 7, which is characterized in that further include：
Terminal according to described search ask to service centre send video data inquiry instruction, so as to the service centre according to
The inquiry instruction searches video data in database.
9. a kind of terminal, which is characterized in that including：
Request reception unit, the searching request for receiving user's transmission；Wherein, described search request include：Search for target；
Data collection unit collects video data for being asked according to described search；
First transmission unit, for described search request and the video data to be sent to service centre, so as to the clothes
Business asks to carry out target identification to the video data according to described search centrally through deep learning network, generates service letter
Breath.
10. a kind of target search tracing system based on video, which is characterized in that including：
Terminal, the searching request for receiving user's transmission；Wherein, described search request include：Search for target；It is searched according to described
Video data is collected in rope request；Described search request and the video data are sent to service centre, so as to the service
It is asked to video data progress target identification, generation information on services according to described search centrally through deep learning network；
Service centre, searching request and video data for receiving terminal transmission；Wherein, described search request include：It searches
Rope target and service request；The service request includes：Target search, target tracking, Address Recognition, circuit recommendation, in real time
At least one of picture；The video data is input to deep learning network trained in advance and carries out target identification, obtains mesh
Mark recognition result；Service processing is carried out to the target identification result according to the service request, obtains corresponding information on services.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810643119.9A CN108764215A (en) | 2018-06-21 | 2018-06-21 | Target search method for tracing, system, service centre and terminal based on video |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810643119.9A CN108764215A (en) | 2018-06-21 | 2018-06-21 | Target search method for tracing, system, service centre and terminal based on video |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108764215A true CN108764215A (en) | 2018-11-06 |
Family
ID=63979748
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810643119.9A Pending CN108764215A (en) | 2018-06-21 | 2018-06-21 | Target search method for tracing, system, service centre and terminal based on video |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108764215A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993163A (en) * | 2019-03-19 | 2019-07-09 | 苏州大学 | A kind of non-rating plate identifying system and its recognition methods based on artificial intelligence |
| CN110675586A (en) * | 2019-09-25 | 2020-01-10 | 捻果科技(深圳)有限公司 | Airport enclosure intrusion monitoring method based on video analysis and deep learning |
| CN111753129A (en) * | 2019-03-26 | 2020-10-09 | 百度在线网络技术（北京）有限公司 | Method, system and terminal equipment for stimulating search based on real-time video content |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8457356B2 (en) * | 2010-10-21 | 2013-06-04 | SET Corporation | Method and system of video object tracking |
| CN105976400A (en) * | 2016-05-10 | 2016-09-28 | 北京旷视科技有限公司 | Object tracking method and device based on neural network model |
| CN106127815A (en) * | 2016-07-21 | 2016-11-16 | 广东工业大学 | A kind of tracking merging convolutional neural networks and system |
| CN106249267A (en) * | 2016-09-30 | 2016-12-21 | 南方科技大学 | Target positioning and tracking method and device |
| CN107808122A (en) * | 2017-09-30 | 2018-03-16 | 中国科学院长春光学精密机械与物理研究所 | Method for tracking target and device |
-
2018
- 2018-06-21 CN CN201810643119.9A patent/CN108764215A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8457356B2 (en) * | 2010-10-21 | 2013-06-04 | SET Corporation | Method and system of video object tracking |
| CN105976400A (en) * | 2016-05-10 | 2016-09-28 | 北京旷视科技有限公司 | Object tracking method and device based on neural network model |
| CN106127815A (en) * | 2016-07-21 | 2016-11-16 | 广东工业大学 | A kind of tracking merging convolutional neural networks and system |
| CN106249267A (en) * | 2016-09-30 | 2016-12-21 | 南方科技大学 | Target positioning and tracking method and device |
| CN107808122A (en) * | 2017-09-30 | 2018-03-16 | 中国科学院长春光学精密机械与物理研究所 | Method for tracking target and device |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109993163A (en) * | 2019-03-19 | 2019-07-09 | 苏州大学 | A kind of non-rating plate identifying system and its recognition methods based on artificial intelligence |
| CN109993163B (en) * | 2019-03-19 | 2023-09-05 | 苏州大学 | Non-tag identification system based on artificial intelligence and identification method thereof |
| CN111753129A (en) * | 2019-03-26 | 2020-10-09 | 百度在线网络技术（北京）有限公司 | Method, system and terminal equipment for stimulating search based on real-time video content |
| CN110675586A (en) * | 2019-09-25 | 2020-01-10 | 捻果科技(深圳)有限公司 | Airport enclosure intrusion monitoring method based on video analysis and deep learning |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10812761B2 (en) | Complex hardware-based system for video surveillance tracking | |
| CN105677793B (en) | The recommended method and device of the foundation of locality database and candidate pick-up point | |
| CN111092938A (en) | Smart city management system based on cloud platform | |
| RU2018137829A (en) | METHOD, DEVICE AND INFORMATION DISPLAY SYSTEM | |
| CN107578487A (en) | A kind of cruising inspection system based on augmented reality smart machine | |
| CN108764215A (en) | Target search method for tracing, system, service centre and terminal based on video | |
| CN104717743B (en) | Mobile terminal method for tracing and system based on signalling analysis | |
| CN107077504A (en) | Supplemental information is asked | |
| CN111435304B (en) | Method and device for generating space unit, storage medium and processor | |
| CN108765610A (en) | Attendance management method for work transmission line scene | |
| CN108132952A (en) | A kind of active searching method and device based on speech recognition | |
| CN104008128A (en) | Automatic case information generating and displaying method | |
| CN111125290B (en) | Intelligent river patrol method and device based on river growth system and storage medium | |
| CN104077360A (en) | GIS (Geographic Information System) real estate valuation based data processing system and method | |
| CN107369123A (en) | Emergent simulated maneuver scheme generation method, device, medium and computer equipment | |
| CN110177255A (en) | A kind of video information dissemination method and system based on case scheduling | |
| CN117391356A (en) | Scheduling method and system for emergency command | |
| CN109889776A (en) | Method for processing video frequency, device, computer installation and computer readable storage medium | |
| CN107222712A (en) | Police command dispatching system and dispatching method | |
| CN109857829A (en) | A kind of geographic information data fusion system | |
| CN108427947A (en) | A kind of image-recognizing method and electronic equipment | |
| CN108335219A (en) | AR social contact methods | |
| CN203204618U (en) | Police standard address database system based on collecting and processing of mobile intelligent terminal | |
| CN111415137A (en) | Violation phenomenon processing method, device, equipment and storage medium | |
| CN107844576B (en) | A kind of environmentally friendly orbit generation method and system of patrolling |
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 | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181106 |
|
| RJ01 | Rejection of invention patent application after publication |