CN110245609A - Pedestrian track generation method, device and readable storage medium storing program for executing - Google Patents
Pedestrian track generation method, device and readable storage medium storing program for executing Download PDFInfo
- Publication number
- CN110245609A CN110245609A CN201910513473.4A CN201910513473A CN110245609A CN 110245609 A CN110245609 A CN 110245609A CN 201910513473 A CN201910513473 A CN 201910513473A CN 110245609 A CN110245609 A CN 110245609A
- Authority
- CN
- China
- Prior art keywords
- pedestrian
- image
- tracking
- track
- template
- 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
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- 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
-
- 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
- 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
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N7/00—Television systems
- H04N7/18—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast
- H04N7/181—Closed-circuit television [CCTV] systems, i.e. systems in which the video signal is not broadcast for receiving images from a plurality of remote sources
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30196—Human being; Person
-
- 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
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30241—Trajectory
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Signal Processing (AREA)
- Human Computer Interaction (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of pedestrian track generation method, device and readable storage medium storing program for executing, comprising the following steps: obtains the pedestrian track information that each camera detection arrives, the pedestrian track information includes pedestrian and image background；Each pedestrian track information is inputted to default neural network model respectively, to obtain the corresponding space time correlation depth characteristic of each pedestrian track information, the space time correlation depth characteristic includes pedestrian's feature, background and pedestrian track；Determine the similarity of each space time correlation depth characteristic；The similarity is greater than or equal to the corresponding pedestrian track information association of space time correlation depth characteristic of preset threshold to generate pedestrian track figure.Because the present invention can obtain space time correlation depth characteristic according to the trace information for including pedestrian's background, matching is carried out across the similarity between camera by space time correlation depth characteristic and generates pedestrian track figure, therefore solves the problems, such as that ken feature difference is larger when the matching of across camera pedestrian track.
Description
Technical field
The present invention relates to field of video monitoring more particularly to a kind of pedestrian track generation methods, device and readable storage
Medium.
Background technique
With carrying forward vigorously for the projects such as safe city, smart city and bright as snow engineering, video monitoring is intelligentized important
Property is increasing.Video monitoring obtains a large amount of pedestrian information, but a large amount of pedestrian track information is not utilized well.
Across camera pedestrian track tracking and matching technique become another video monitoring intelligence after face recognition technology
Hot spots for development can be changed, play very important role.Across camera pedestrian track tracking and matching technique mainly utilize pedestrian
Detection, the tracking of single camera pedestrian track, across camera pedestrian track matching technique carry out pedestrian track association.But it is existing across
Camera pedestrian track tracking is not right simultaneously due to being matched across between camera according to pedestrian's characteristic similarity
The features such as space-time, background and the prospect in tracking frame that pedestrian occurs match, when across camera pedestrian track being caused to match
Ken feature difference is larger, to reduce the matched accuracy of across camera pedestrian track.
Above content is only used to facilitate the understanding of the technical scheme, and is not represented and is recognized that above content is existing skill
Art.
Summary of the invention
The main purpose of the present invention is to provide a kind of pedestrian track generation method, device and readable storage medium storing program for executing, purports
When solving the problems, such as the matching of across camera pedestrian track, ken feature difference is larger causes the matched accuracy of pedestrian track low.
In order to achieve the above object, the present invention provides a kind of pedestrian track generation method, the pedestrian track generation method
The following steps are included:
The pedestrian track information that each camera detection arrives is obtained, the pedestrian track information includes pedestrian and image back
Scape；
Each pedestrian track information is inputted to default neural network model respectively, to obtain each pedestrian track
The corresponding space time correlation depth characteristic of information, the space time correlation depth characteristic include pedestrian's feature, background and pedestrian's rail
Mark；
Determine the similarity of each space time correlation depth characteristic；
The corresponding pedestrian track information of the space time correlation depth characteristic that the similarity is greater than or equal to preset threshold is closed
Connection, to generate pedestrian track figure.
Preferably, it is described obtain each camera detection to pedestrian track information the step of include:
Obtain the collected tracking frame sequence of each camera；
Determine the sampling tracking frame in each tracking frame sequence；
Pedestrian's template library is updated according to each sampling tracking frame；
Pedestrian track frame is generated according to updated pedestrian's template library；
Obtain the newest pedestrian tracking template in the pedestrian track frame；
According to the corresponding pedestrian track information of each newest pedestrian tracking template generation.
Preferably, the step of pedestrian track information corresponding according to each newest pedestrian tracking template generation is wrapped
It includes:
The pedestrian image store path of each newest pedestrian tracking template is obtained, and according to each pedestrian image
Store path obtains corresponding newest pedestrian image；
According to each newest pedestrian image to the tracking frame in the tracking frame sequence in addition to the sampling tracking frame
It is filtered, and generates the corresponding pedestrian track information of each newest pedestrian image.
Preferably, described the step of updating pedestrian's template library according to each sampling tracking frame, includes:
Obtain multiple first pedestrian images in each sampling tracking frame；
Obtain corresponding second pedestrian image of all pedestrian tracking templates in pedestrian's template library；
Obtain the characteristic information of each first pedestrian image and the characteristic information of each second pedestrian image；
Determine the characteristic information of each first pedestrian image and the characteristic information of each second pedestrian image
Similarity；
Pedestrian's template library is updated according to each similarity.
Preferably, described the step of updating pedestrian's template library according to each similarity, includes:
The similarity for judging whether there is first pedestrian image and multiple second pedestrian images is greater than or equal to
Preset threshold；
It is pre- determining to be greater than or equal to there are the similarity of second pedestrian image and multiple first pedestrian images
If when threshold value, obtaining highest second pedestrian image of similarity；
Obtain pedestrian's mark in the corresponding pedestrian tracking template of highest second pedestrian image of the similarity；
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to described first
Pedestrian's template library described in the corresponding pedestrian tracking template renewal of pedestrian image.
Preferably, the similarity for judging whether there is first pedestrian image Yu multiple second pedestrian images
After the step of more than or equal to preset threshold, further includes:
In judgement, there is no the similarities of first pedestrian image and multiple second pedestrian images to be greater than or equal to
When preset threshold, pedestrian's mark is generated by preset algorithm；
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to described first
Pedestrian's template library described in the corresponding pedestrian tracking template renewal of pedestrian image.
Preferably, described that the step for generating the corresponding pedestrian tracking template of first pedestrian image is identified according to the pedestrian
Suddenly include:
Obtain the characteristic information of first pedestrian image；
Obtain the storage road of location information of first pedestrian image in the tracking frame, first pedestrian image
Diameter, the camera identification for generating the sampling tracking frame；
According to the pedestrian mark, generate it is described sampling tracking frame camera identification, sample tracking frame the generation time,
Location information, the characteristic information of the first pedestrian image and depositing for first pedestrian image in corresponding sampling tracking frame
Store up the pedestrian tracking template of the first pedestrian image described in coordinates measurement, wherein the generation time of the sampling tracking frame is described
Camera collects the time obtained when tracking frame sequence.
Preferably, after the step of pedestrian tracking template for generating first pedestrian image, further includes:
It is newest pedestrian tracking template by the corresponding pedestrian tracking template-setup of first pedestrian image.
Preferably, it is described by the corresponding pedestrian tracking template-setup of first pedestrian image be newest pedestrian tracking template
The step of after, further includes:
Obtain in each second pedestrian image be greater than with the similarity of the characteristic information of first pedestrian image or
It equal to default similarity, and is the pedestrian tracking template of newest pedestrian tracking template；
It will whether be that the value of newest pedestrian tracking template is set as no in the pedestrian tracking template.
Preferably, whether described will be that the value of newest pedestrian tracking template is set as no step in the pedestrian tracking template
After rapid, comprising:
It obtains and generates the pedestrian tracking template of time and current time difference more than or equal to preset threshold in template library；
Delete the pedestrian tracking template that the difference is greater than or equal to preset threshold.
In addition, to achieve the above object, the present invention also provides a kind of pedestrian track generating means, the pedestrian track is generated
Device includes that the pedestrian track that can run on the memory and on the processor of processor, memory and being stored in generates
Program, the pedestrian track generate the step that pedestrian track generation method as described above is realized when program is executed by the processor
Suddenly.
In addition, to achieve the above object, the present invention also provides a kind of readable storage medium storing program for executing, being deposited on the readable storage medium storing program for executing
It contains pedestrian track and generates program, the pedestrian track generates when program is executed by processor and realizes pedestrian track as described above
The step of generation method.
Pedestrian track generation method, device and readable storage medium storing program for executing provided by the invention, firstly, obtaining each camera
Then each pedestrian track information is inputted default neural network model by the pedestrian track information detected respectively, with
To the corresponding space time correlation depth characteristic of each pedestrian track information, then, determine each space time correlation depth
The similarity of feature, finally, the similarity to be greater than or equal to the corresponding pedestrian of space time correlation depth characteristic of preset threshold
Trace information association, to generate pedestrian track figure.Because the present invention can obtain the trace information of pedestrian, root from each camera
Space time correlation depth characteristic is obtained according to trace information, across the similarity progress rail between camera by space time correlation depth characteristic
Mark matching generates pedestrian track figure, so that ken feature difference is larger when solving the matching of across camera pedestrian track, across camera shooting
The low problem of the accuracy of head's path matching.
Detailed description of the invention
Detailed description of the invention is used to provide further understanding of the present invention, and constitutes part of specification, with the present invention
Embodiment be used to explain the present invention together, be not construed as limiting the invention.In the accompanying drawings:
Fig. 1 is the hardware structural diagram for the pedestrian track generating means that the embodiment of the present invention is related to；
Fig. 2 is the flow diagram of pedestrian track generation method first embodiment of the present invention；
Fig. 3 is the flow diagram of pedestrian track generation method second embodiment of the present invention；
Fig. 4 is the flow diagram of pedestrian track generation method 3rd embodiment of the present invention；
Fig. 5 is the flow diagram of pedestrian track generation method fourth embodiment of the present invention；
Fig. 6 is the flow diagram of the 5th embodiment of pedestrian track generation method of the present invention；
Fig. 7 is the flow diagram of pedestrian track generation method sixth embodiment of the present invention；
Fig. 8 is the flow diagram of the 7th embodiment of pedestrian track generation method of the present invention.
The embodiments will be further described with reference to the accompanying drawings for the realization, the function and the advantages of the object of the present invention.
Specific embodiment
It should be appreciated that the specific embodiments described herein are merely illustrative of the present invention, it is not intended to limit the present invention.
Due to existing pedestrian track generation method algorithm accuracy and calculation in terms of pedestrian track generation in single camera
Method performance is lower, and do not have in terms of across camera pedestrian track matching using deep learning network extract pedestrian image in when
The features such as empty, background and prospect, cause across the ken feature difference larger, so that it is matched to substantially reduce across camera pedestrian track
Accuracy.
The present invention provides a solution, firstly, the pedestrian track information that each camera detection arrives is obtained, then,
Each pedestrian track information is inputted to default neural network model respectively, to obtain each pedestrian track information
Corresponding space time correlation depth characteristic, then, the similarity of each space time correlation depth characteristic is determined, finally, by the phase
It is greater than or equal to the corresponding pedestrian track information association of space time correlation depth characteristic of preset threshold, like degree to generate pedestrian track
Figure.Because the present invention can obtain the trace information of pedestrian from each camera, space time correlation depth is obtained according to trace information
Feature generates pedestrian track figure across path matching is carried out by the similarity of space time correlation depth characteristic between camera, thus
Ken feature difference is larger when solving the matching of across camera pedestrian track, and across the camera matched accuracy of pedestrian track is low
Problem.
As shown in Figure 1, Fig. 1 is that the embodiment of the present invention is related to the hardware structural diagram of device.
Referring to Fig.1, the apparatus may include processor 1001, such as CPU, memory 1002, communication bus 1003, nets
Network interface 1004.Wherein, communication bus 1003 is for realizing the connection communication between each building block in the device.Network interface
1004 may include optionally standard wireline interface and wireless interface (such as WI-FI interface).Memory 1002 can be high speed
RAM memory is also possible to stable memory (non-volatile memory), such as magnetic disk storage.Memory 1002
It optionally can also be the storage device independently of aforementioned processor 1001.As shown in Figure 1, as a kind of computer storage medium
Memory 1002 in may include that operating system, network communication module and pedestrian track generate program.
Optionally, described device can also include camera, RF (Radio Frequency, radio frequency) circuit, sensor,
Voicefrequency circuit, WiFi module etc..Wherein, sensor such as optical sensor, motion sensor and other sensors.Specifically
Ground, optical sensor may include ambient light sensor and proximity sensor, wherein ambient light sensor can be according to the bright of ambient light
The brightness of display screen secretly is adjusted, proximity sensor can close display screen and/or backlight when intelligent terminal is moved in one's ear.
As a kind of motion sensor, gravity accelerometer can detect all directions on (generally three axis) acceleration it is big
It is small, it can detect that size and the direction of gravity when static, can be used to identify the application of intelligent terminal posture, (for example horizontal/vertical screen is cut
Change, dependent game, magnetometer pose calibrating), Vibration identification correlation function (such as pedometer, tap) etc.；Certainly, intelligent terminal
It can also configure the other sensors such as gyroscope, barometer, hygrometer, thermometer, infrared sensor, details are not described herein.
It will be understood by those skilled in the art that the restriction of the not structure twin installation of apparatus structure shown in Fig. 1, can wrap
It includes than illustrating more or fewer components, perhaps combines certain components or different component layouts.
In hardware involved in device shown in Fig. 1, network interface 1004 can be used for uploading collected tracking frame；
And processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, and execute following operation:
The pedestrian track information that each camera detection arrives is obtained, the pedestrian track information includes pedestrian and image back
Scape；
Each pedestrian track information is inputted to default neural network model respectively, to obtain each pedestrian track
The corresponding space time correlation depth characteristic of information, the space time correlation depth characteristic include pedestrian's feature, background and pedestrian's rail
Mark；
Determine the similarity of each space time correlation depth characteristic；
The corresponding pedestrian track information of the space time correlation depth characteristic that the similarity is greater than or equal to preset threshold is closed
Connection, to generate pedestrian track figure.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
Obtain the collected tracking frame sequence of each camera；
Determine the sampling tracking frame in each tracking frame sequence；
Pedestrian's template library is updated according to each sampling tracking frame；
Pedestrian track frame is generated according to updated pedestrian's template library；
Obtain the newest pedestrian tracking template in the pedestrian track frame；
According to the corresponding pedestrian track information of each newest pedestrian tracking template generation.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
The pedestrian image store path of each newest pedestrian tracking template is obtained, and according to each pedestrian image
Store path obtains corresponding newest pedestrian image；
According to each newest pedestrian image to the tracking frame in the tracking frame sequence in addition to the sampling tracking frame
It is filtered, and generates the corresponding pedestrian track information of each newest pedestrian image.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
Obtain multiple first pedestrian images in each sampling tracking frame；
Obtain corresponding second pedestrian image of all pedestrian tracking templates in pedestrian's template library；
Obtain the characteristic information of each first pedestrian image and the characteristic information of each second pedestrian image；
Determine the characteristic information of each first pedestrian image and the characteristic information of each second pedestrian image
Similarity；
Pedestrian's template library is updated according to each similarity.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
The similarity for judging whether there is first pedestrian image and multiple second pedestrian images is greater than or equal to
Preset threshold；
It is pre- determining to be greater than or equal to there are the similarity of second pedestrian image and multiple first pedestrian images
If when threshold value, obtaining highest second pedestrian image of similarity；
Obtain pedestrian's mark in the corresponding pedestrian tracking template of highest second pedestrian image of the similarity；
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to described first
Pedestrian's template library described in the corresponding pedestrian tracking template renewal of pedestrian image.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
In judgement, there is no the similarities of first pedestrian image and multiple second pedestrian images to be greater than or equal to
When preset threshold, pedestrian's mark is generated by preset algorithm；
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to described first
Pedestrian's template library described in the corresponding pedestrian tracking template renewal of pedestrian image.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
Obtain the characteristic information of first pedestrian image；
Obtain the storage road of location information of first pedestrian image in the tracking frame, first pedestrian image
Diameter, the camera identification for generating the sampling tracking frame；
According to the pedestrian mark, generate it is described sampling tracking frame camera identification, sample tracking frame the generation time,
Location information, the characteristic information of the first pedestrian image and depositing for first pedestrian image in corresponding sampling tracking frame
Store up the pedestrian tracking template of the first pedestrian image described in coordinates measurement, wherein the generation time of the sampling tracking frame is described
Camera collects the time obtained when tracking frame sequence.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
It is newest pedestrian tracking template by the corresponding pedestrian tracking template-setup of first pedestrian image.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
Obtain in each second pedestrian image be greater than with the similarity of the characteristic information of first pedestrian image or
It equal to default similarity, and is the pedestrian tracking template of newest pedestrian tracking template；
It will whether be that the value of newest pedestrian tracking template is set as no in the pedestrian tracking template.
Further, processor 1001 can be used for that the pedestrian track stored in memory 1002 is called to generate program, also
Execute following operation:
It obtains and generates the pedestrian tracking template of time and current time difference more than or equal to preset threshold in template library；
Delete the pedestrian tracking template that the difference is greater than or equal to preset threshold.
It is the first embodiment of pedestrian track generation method of the present invention, the pedestrian track generation method referring to Fig. 2, Fig. 2
Include:
Step S10, the pedestrian track information that each camera detection arrives is obtained；
Pedestrian track generation method provided by the invention is mainly used for across the trajectory diagram for generating pedestrian between camera.The present invention
The terminal that the pedestrian track generation method of offer is related to includes but is not limited to mobile phone, tablet computer and computer etc., in the terminal
It is pre-loaded with relevant application system.
Technical solution provided by the invention, the pedestrian track information are to pass through pretreated tracking frame sequence, represent row
The continuous path information of people includes pedestrian and image background in tracking frame.Due to existing across camera pedestrian track match party
Method does not have to obtain when across camera the space-time of the pedestrian image that different cameras obtain, background and preceding using deep learning network
The feature of scape etc., therefore while each camera obtains pedestrian image, retain the background and/or prospect letter in tracking frame
Breath.Camera collects tracking frame sequence first, sampling tracking frame is then determined from the tracking frame sequence, then obtain the sampling
One or more pedestrian images in tracking frame, then the characteristic information of each pedestrian image is obtained, believe further according to multiple features
Breath updates pedestrian's template library, determines pedestrian track frame further according to updated pedestrian's template library, then obtain pedestrian track frame
In multiple newest pedestrian tracking templates, finally obtain multiple corresponding pedestrian tracks according to multiple pedestrian tracking templates
Figure.It is understood that repeating the above steps after camera gets tracking frame sequence and obtaining pedestrian track figure.
Step S20, default neural network model is inputted respectively by each pedestrian track information, it is each described to obtain
The corresponding space time correlation depth characteristic of pedestrian track information；
In technical solution provided in this embodiment, the space time correlation depth characteristic includes pedestrian's feature, background and pedestrian
Track.The default neural network model is space time correlation ConvLSTM neural network model, for obtaining pedestrian track information
Corresponding space time correlation depth characteristic.By the pedestrian track information input space time correlation ConvLSTM neural network model
Afterwards, it is special to extract pedestrian's feature, background and/or prospect in pedestrian track information for space time correlation ConvLSTM neural network model
It is deep to generate space time correlation further according to pedestrian's feature, background and/or foreground features and pedestrian track for sign and pedestrian track
Spend feature.
Step S30, the similarity of each space time correlation depth characteristic is determined；
In technical solution provided in this embodiment, the space time correlation depth characteristic of the pedestrian under each camera is got
Afterwards, the similarity calculation algorithm for obtaining space time correlation depth characteristic, calculates the similarity of each space time correlation depth characteristic.
Step S40, the similarity is greater than or equal to the corresponding pedestrian's rail of space time correlation depth characteristic of preset threshold
Mark information association, to generate pedestrian track figure.
In technical solution provided in this embodiment, the preset threshold can be 60%, it is to be understood that in order to improve
Across the camera matched accuracy of pedestrian track, preset threshold can be correspondinglyd increase is 70% or 80%.Across general between camera
Similarity is greater than or equal to the corresponding pedestrian track information association of space time correlation depth characteristic of preset threshold, after obtaining association
The pedestrian track information of multiple space time correlation depth characteristics connects software by track and connects multiple pedestrian track information,
Obtain the pedestrian track figure.It is understood that the pedestrian track information can also be according to the life of space time correlation depth information
It is generated at the time, i.e., similarity is greater than to the corresponding pedestrian track information association of space time correlation depth characteristic of preset threshold, so
The generation time of associated pedestrian track information is obtained afterwards, then by the associated pedestrian track information by the generation time
Sequence is matched, to generate pedestrian track figure.Due to being easy in the trajectory diagram of across camera generation pedestrian by pedestrian track
Information connection error, generation does not meet actual pedestrian track figure, therefore similarity is greater than to the space time correlation depth of preset threshold
The corresponding pedestrian track information storage of feature is associated into the middle table of database, obtains associated pedestrian track information
The time is generated, software is connected by track and connects multiple pedestrian tracks letters by the generation time sequencing of pedestrian track information
Breath, obtains the pedestrian track figure.By the generation time connection pedestrian track information of pedestrian track information to generate across camera
Pedestrian track figure, to avoid the pedestrian track figure of generation error.
It is understood that after generating pedestrian track figure preservation can also be associated with corresponding pedestrian.It goes generating
After people's trajectory diagram, obtain the corresponding pedestrian's mark of each pedestrian, then will each pedestrian identify with it is corresponding described
Pedestrian track figure is associated preservation.Specifically, the store path storage pedestrian track figure deposited is identified to the pedestrian
In corresponding database table.It is convenient so directly to transfer the pedestrian's from database when inquiring or tracking some specific pedestrian
Pedestrian track figure saves the time required when generating pedestrian track figure.
The present invention can obtain the trace information of pedestrian from each camera, obtain space time correlation depth according to trace information
Feature is spent, generates pedestrian track figure across path matching is carried out by the similarity of space time correlation depth characteristic between camera, from
And ken feature difference when reducing the matching of across camera pedestrian track, it is matched accurate to improve across camera pedestrian track
Property.
Further, it is the second embodiment of pedestrian track generation method of the present invention referring to Fig. 3, Fig. 3, is based on above-mentioned implementation
Example, the step S10, comprising:
Step S11, each collected tracking frame sequence of camera is obtained；
Step S12, the sampling tracking frame in each tracking frame sequence is determined；
Step S13, pedestrian's template library is updated according to each sampling tracking frame；
Step S14, pedestrian track frame is generated according to updated pedestrian's template library；
Step S15, the newest pedestrian tracking template in the pedestrian track frame is obtained；
Step S16, according to the corresponding pedestrian track information of each newest pedestrian tracking template generation.
In technical solution provided in this embodiment, the tracking frame sequence includes multiple tracking frames, the sampling tracking frame
For the frame in multiple tracking frames of the tracking frame sequence, wherein the sampling tracking frame can be the tracking frame sequence
In any one frame, the sampling tracking frame can also obtain at random by algorithm.Pedestrian's template library can according to sampling with
Track frame just updates, and can also delete outmoded pedestrian tracking template over time.The pedestrian track frame is by pedestrian
Template library filling, i.e., the updated information of pedestrian's template library will be updated to the pedestrian track frame.Due in tracking frame sequence
There are multiple tracking frames, when can increase the expense and pedestrian track map generalization of system resource to the processing of each tracking frame
Between.Therefore after camera gets tracking frame sequence, a frame therein is only obtained as sampling tracking frame, if tracking frame sequence has
When multiple, it is determined that then the sampling tracking frame in each tracking frame sequence is updated according to each sampling tracking frame
Pedestrian's template library generates pedestrian track frame further according to updated pedestrian's template library, then obtains the pedestrian track
In frame in pedestrian tracking template whether be newest template value be or 1 pedestrian tracking template, it is to be understood that most
New pedestrian's trace template may have one or more.It is finally corresponding according to one or more newest pedestrian tracking template generations
Pedestrian track information.
The present invention obtains pedestrian track figure by sampling tracking frame, to reduce the expense and pedestrian's rail of system resource
The mark map generalization time.
Further, referring to Fig. 4, Fig. 4 is the 3rd embodiment of pedestrian track generation method of the present invention, in above-mentioned Fig. 3 institute
On the basis of the embodiment shown, the step S16, comprising:
Step S161, the pedestrian image store path of each newest pedestrian tracking template is obtained, and according to each institute
It states pedestrian image store path and obtains corresponding newest pedestrian image；
Step S162, according to each newest pedestrian image in the tracking frame sequence in addition to the sampling tracking frame
Tracking frame be filtered, and generate the corresponding pedestrian track information of each newest pedestrian image.
In technical solution provided in this embodiment, the pedestrian image store path is obtained from sampling tracking frame for storing
The one or more pedestrian images taken.The filtering processing was the pedestrian filtered out outside people from nominated bank.In order to be generated across camera
Pedestrian track information needs the pedestrian track information that the pedestrian traced into is first generated in single camera.Obtain it is each it is described most
The pedestrian image store path of new pedestrian's trace template, and it is corresponding newest according to each pedestrian image store path acquisition
Pedestrian image.Then according to each newest pedestrian image to it is described tracking frame sequence in except it is described sampling tracking frame in addition to
Track frame is filtered, and generates the corresponding pedestrian track information of each newest pedestrian image.Specifically, in order to more preferable
Ground understands this programme, filtering processing of illustrating by taking single newest pedestrian image as an example below.After getting newest pedestrian image, obtain
The corresponding tracking frame sequence of the newest pedestrian image, it is assumed that the tracking frame sequence includes 1 to 10 frames, and sampling tracking frame is the 5th frame,
What is be then filtered is tracked as 1 to 4,6 to 10 frames, i.e. the 5th frame in the tracking frame sequence is not filtered.Then from
First frame starts, which is from left to right slided from the upper left corner of first frame tracking frame, coasting distance and cunning
Scanning frequency degree can be preset by system, calculate the image and the phase of the newest pedestrian image slided within the scope of frame during sliding
Scheme until sliding complete tracking frame if calculating during sliding to the similarity of the image slided in frame and newest pedestrian like degree
The similarity of picture is greater than preset threshold, then obtains the information such as location information, background and prospect of the pedestrian in the tracking frame.So
The second frame is filtered afterwards, is repeated the above steps, until completing to be filtered to the tracking frame sequence.Last basis obtains
The information such as multiple location informations, background and the prospect got generate pedestrian track information.
According to each pedestrian image to corresponding row in the tracking frame in corresponding tracking frame sequence in addition to sampling tracking frame
People's image is filtered, and obtains the pedestrian track information of each pedestrian, because the pedestrian track information includes institute for multiframe
The tracking frame sequence of pedestrian and background is stated, so that the pedestrian track information generated is more three-dimensional, according to pedestrian track information
The space time correlation depth characteristic of generation is more accurate.
Further, referring to Fig. 5, Fig. 5 is the fourth embodiment of pedestrian track generation method of the present invention, in above-mentioned Fig. 3 institute
On the basis of the embodiment shown, the step S13, comprising:
Step S131, multiple first pedestrian images in each sampling tracking frame are obtained；
Step S132, corresponding second pedestrian image of all pedestrian tracking templates in pedestrian's template library is obtained；
Step S133, obtain each first pedestrian image characteristic information and each second pedestrian image
Characteristic information；
Step S134, the characteristic information of each first pedestrian image and the spy of each second pedestrian image are determined
The similarity of reference breath；
Step S135, pedestrian's template library is updated according to each similarity.
In technical solution provided in this embodiment, the characteristic information is LOMO characteristic information, the LOMO characteristic information
It is extracted by default feature extraction algorithm, that is, LOMO feature extraction algorithm.Pedestrian's template library occurs according in sampling tracking frame
Pedestrian image be updated.Specifically, multiple first pedestrian images in each sampling tracking frame are obtained first, then
All pedestrian tracking templates in pedestrian's template library are obtained, the pedestrian image store path of all pedestrian tracking templates is passed through
Obtain corresponding second pedestrian image.Then the feature of multiple first pedestrian images is extracted by LOMO feature extraction algorithm
Information obtains the corresponding characteristic information of all second pedestrian images by all pedestrian tracking templates.Due to pedestrian with
The angle occurred in track frame is all different, therefore again by the characteristic information of each the first pedestrian image and all second pedestrians
Similarity is calculated between the characteristic information of image, is determined as the existing pedestrian when similarity is greater than or equal to preset threshold,
It is less than preset threshold in similarity or there is no be determined as new pedestrian when having similarity.It is understood that because being each
Similarity is calculated between the characteristic information of first pedestrian image and the characteristic information of all second pedestrian images, it is thus possible to
Will appear the first pedestrian image for participating in calculating, there are similarities with the one or more the second pedestrian images.This depends on the pedestrian
The frequency occurred in previous sampling tracking frame.
By the characteristic information of all second pedestrian images in the characteristic information of the first pedestrian image and pedestrian's template library into
Row matching, to improve accuracy when characteristic information matching.
Further, referring to Fig. 6, Fig. 6 is the 5th embodiment of pedestrian track generation method of the present invention, in above-mentioned Fig. 5 institute
On the basis of the embodiment shown, the step S135 includes:
Step S1351, the similarity of first pedestrian image Yu multiple second pedestrian images is judged whether there is
More than or equal to preset threshold；
Step S1352, determining that there are the similarity of second pedestrian image and multiple first pedestrian images is big
When preset threshold, highest second pedestrian image of similarity is obtained；
Step S1353, the pedestrian in the corresponding pedestrian tracking template of highest second pedestrian image of similarity is obtained
Mark；
Step S1354, it is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and root
According to pedestrian's template library described in the corresponding pedestrian tracking template renewal of first pedestrian image.
In technical solution provided in this embodiment, the pedestrian is identified for identifying the pedestrian of uniqueness, if recognize
Pedestrian image does not have corresponding pedestrian in pedestrian's trace template, then generates new pedestrian ID and mark as the pedestrian of the pedestrian
Know.If there are corresponding pedestrians in pedestrian's trace template for the pedestrian image recognized, the pedestrian of pedestrian tracking template is obtained
It identifies and is identified as the pedestrian of the pedestrian image.Since pedestrian may repeatedly appear in sampling tracking frame, therefore pedestrian's template library
In the corresponding pedestrian tracking template of the pedestrian can also exist it is multiple.Before obtaining pedestrian's mark of the first pedestrian image, sentence first
It is disconnected to be greater than or equal to preset threshold with the presence or absence of the similarity of first pedestrian image and multiple second pedestrian images,
When determining to be greater than or equal to preset threshold there are the similarity of second pedestrian image and multiple first pedestrian images, obtain
Highest second pedestrian image of similarity is taken, the corresponding pedestrian tracking of highest second pedestrian image of the similarity is then obtained
Pedestrian's mark in template, the pedestrian of highest second pedestrian image of the similarity is finally identified be used as first pedestrian
The pedestrian of image identifies, identified according to the pedestrian of first pedestrian image generate the corresponding pedestrian of first pedestrian image with
Track template, and pedestrian's template library according to first pedestrian image corresponding pedestrian tracking template renewal.
In other embodiments, based on the basis of above-mentioned embodiment shown in fig. 6, the step of the step S1351 after,
Further include:
Step S1355, in the similarity for determining that first pedestrian image and multiple second pedestrian images is not present
When more than or equal to preset threshold, pedestrian's mark is generated by preset algorithm；
Step S1356, it is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and root
According to pedestrian's template library described in the corresponding pedestrian tracking template renewal of first pedestrian image.
In technical solution provided in this embodiment, determining that there is no first pedestrian image and multiple second rows
When the similarity of people's image is greater than or equal to preset threshold, pedestrian's mark, root are generated by preset algorithm (such as UUID algorithm)
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to first pedestrian image pair
Pedestrian's template library described in the pedestrian tracking template renewal answered.
Pedestrian's template library is updated by sampling tracking frame after determining sampling tracking frame, makes to determine that sampling tracks next time
Whether match information in the presence of pedestrian's template library is more acurrate by the pedestrian occurred in frame.
It further, is the sixth embodiment of pedestrian track generation method of the present invention referring to Fig. 7, Fig. 7, it is described according to
Pedestrian identifies the step of generating first pedestrian image corresponding pedestrian tracking template and includes:
Step S50, the characteristic information of first pedestrian image is obtained；
Step S60, location information in the tracking frame of first pedestrian image, first pedestrian image are obtained
Store path, generate it is described sampling tracking frame camera identification；
Step S70, it is identified according to the pedestrian, generates the camera identification for sampling tracking frame, samples tracking frame
Generate time, the location information in corresponding sampling tracking frame, the characteristic information of the first pedestrian image and the first row
The store path of people's image generates the pedestrian tracking template of first pedestrian image, wherein the generation of the sampling tracking frame
Time is that the camera collects the time obtained when tracking frame sequence.
In technical solution provided in this embodiment, the pedestrian is identified as pedestrian ID, and the camera identification is camera
ID.In order to facilitate the pedestrian image occurred in each tracking frame is tracked, when extracting one or more pedestrian images, generating should
The corresponding pedestrian tracking template of one or more pedestrian images.
In order to better understand this programme, citing illustrates the process for generating pedestrian tracking template below.Camera C acquisition row
Human feelings condition generates tracking frame A1, carries out pedestrian detection to tracking frame A1 using pedestrian detection algorithm, obtains pedestrian image set B,
If detect 4 pedestrians, set B includes B1, B2, B3, B4, calculates pedestrian image set B using LOMO feature extraction algorithm
In each pedestrian image LOMO feature.Pedestrian is distributed to the corresponding pedestrian of each pedestrian image in pedestrian image set B again
ID, and to each pedestrian according to format (pedestrian ID, camera ID, pedestrian's time of occurrence, pedestrian's appearance position, the LOMO of pedestrian
Feature vector, if as the mark of newest trace template, the name path of pedestrian image) in the corresponding storage table of database
It is middle to save a record.All records constitute the corresponding initial pedestrian's trace template library TA1 of tracking frame A1.Camera C is default
Pedestrian's situation is acquired after interval time, is generated tracking frame A2, is repeated the above steps to obtain pedestrian tracking template library TA2.
It should be noted that comparing tracking frame A2 by circulation when distributing pedestrian ID for the pedestrian image of tracking frame A2
Pedestrian image and tracking frame A1 each pedestrian image, pedestrian characteristic similarity be more than preset threshold when, be judged as
The pedestrian occurred, using the ID of the pedestrian of the appearance as the ID of the pedestrian image of previous cycle.If without spy after circulation
When levying similarity more than preset threshold, then new pedestrian ID is distributed.
In other embodiments, based on the basis of above-mentioned embodiment shown in Fig. 7, the step of the step S70 after, also
Include:
It step S80, is newest pedestrian tracking template by the corresponding pedestrian tracking template-setup of first pedestrian image.
In technical solution provided in this embodiment, one of field in the pedestrian tracking template be whether be newest
Template, the value of the field can be 1 or 0, be also possible to or "No", 1 or "Yes" represent the pedestrian tracking template as newest pedestrian
Trace template, 0 or "No" to represent the pedestrian tracking template be not newest pedestrian tracking template.
In other embodiments, after the step of step S80, further includes:
Step S90, it obtains similar to the characteristic information of first pedestrian image in each second pedestrian image
Degree is greater than or equal to default similarity, and is the pedestrian tracking template of newest pedestrian tracking template；
It step S100, will whether be that the value of newest pedestrian tracking template is set as no in the pedestrian tracking template.
In technical solution provided in this embodiment, since the pedestrian tracking template of highest second pedestrian image of similarity can
It can not be newest pedestrian tracking template, therefore need to judge the corresponding pedestrian tracking mould of highest second pedestrian image of the similarity
The whether newest pedestrian tracking template of plate；It is in the corresponding pedestrian tracking template of judgement highest second pedestrian image of similarity
It will whether be that the value of newest pedestrian tracking template is set as no in the pedestrian tracking template when newest pedestrian tracking template；?
When determining that the corresponding pedestrian tracking template of highest second pedestrian image of the similarity is not newest pedestrian tracking template, according to
The second pedestrian picture obtains similarity and is greater than or equal to preset threshold and is the pedestrian tracking mould of newest pedestrian tracking template
Whether plate will be that the value of newest pedestrian tracking template is set as no in the pedestrian tracking template.
The first pedestrian image pedestrian tracking template-setup be newest pedestrian tracking template after, by before the pedestrian most
New pedestrian's trace template cancel it is newest, to avoid in pedestrian's template library that there are multiple newest pedestrian tracking templates by some pedestrian
Situation.
Further, it is the 7th embodiment of pedestrian track generation method of the present invention referring to Fig. 8, Fig. 8, is based on above-mentioned Fig. 7
Shown on the basis of embodiment, after the step S100, further includes::
Step S110, obtain generated in template library time and current time difference be greater than or equal to the pedestrian of preset threshold with
Track template；
Step S120, the pedestrian tracking template that the difference is greater than or equal to preset threshold is deleted.
In technical solution provided in this embodiment, the trace template frame include the corresponding all pedestrians of the pedestrian with
Track template.It tracks in frame sequence in order to prevent and outmoded pedestrian tracking template occurs, so that the shadow when generating pedestrian track information
The accuracy of the pedestrian track information is rung, therefore outmoded pedestrian tracking template should be deleted.It is most freshly harvested to obtain the camera
The generation time of tracking frame obtains pedestrian's mark of each pedestrian image, and it is each to identify acquisition according to each pedestrian
The corresponding trace template frame of the pedestrian, obtains all pedestrian tracking templates in the trace template frame of each pedestrian
The generation time, determine the difference of the generation time for generating time and the tracking frame of all pedestrian tracking templates respectively
Value is then determined as outmoded template when the difference is greater than or equal to preset threshold, finally deletes the difference and be greater than or equal to
The pedestrian tracking template of preset threshold.The pedestrian tracking template that the difference is less than preset threshold then retains.
That the pedestrian is deleted before generating pedestrian track information is judged as outmoded pedestrian tracking template, to improve
The accuracy of pedestrian track information.
To achieve the above object, the present invention also provides a kind of pedestrian track generating means, the pedestrian track generating means
Including processor, memory and it is stored in the pedestrian track that can be run on the memory and on the processor generation journey
Sequence, the pedestrian track generate the step that pedestrian track generation method as described above is realized when program is executed by the processor
Suddenly.
To achieve the above object, it the present invention also provides a kind of readable storage medium storing program for executing, is stored on the readable storage medium storing program for executing
Pedestrian track generates program, and the pedestrian track generates and realizes that pedestrian track as described above generates when program is executed by processor
The step of method.
It should be noted that, in this document, the terms "include", "comprise" or its any other variant are intended to non-row
His property includes, so that the process, method, article or the system that include a series of elements not only include those elements, and
And further include other elements that are not explicitly listed, or further include for this process, method, article or system institute it is intrinsic
Element.In the absence of more restrictions, the element limited by sentence "including a ...", it is not excluded that including being somebody's turn to do
There is also other identical elements in the process, method of element, article or system.
The serial number of the above embodiments of the invention is only for description, does not represent the advantages or disadvantages of the embodiments.
Through the above description of the embodiments, those skilled in the art can be understood that above-described embodiment side
Method can be realized by means of software and necessary general hardware platform, naturally it is also possible to by hardware, but in many cases
The former is more preferably embodiment.Based on this understanding, technical solution of the present invention substantially in other words does the prior art
The part contributed out can be embodied in the form of software products, which is stored in one as described above
In storage medium (such as ROM/RAM, magnetic disk, CD), including some instructions are used so that a terminal device (can be TV
Machine, mobile phone, computer, device, air conditioner or network equipment etc.) execute method described in each embodiment of the present invention.
The above is only a preferred embodiment of the present invention, is not intended to limit the scope of the invention, all to utilize this hair
Equivalent structure or equivalent flow shift made by bright specification and accompanying drawing content is applied directly or indirectly in other relevant skills
Art field, is included within the scope of the present invention.
Claims (12)
1. a kind of pedestrian track generation method, which is characterized in that the pedestrian track generation method the following steps are included:
The pedestrian track information that each camera detection arrives is obtained, the pedestrian track information includes pedestrian and image background；
Each pedestrian track information is inputted to default neural network model respectively, to obtain each pedestrian track information
The corresponding space time correlation depth characteristic of information, the space time correlation depth characteristic include pedestrian's feature, background and pedestrian track；
Determine the similarity of each space time correlation depth characteristic；
The similarity is greater than or equal to the corresponding pedestrian track information association of space time correlation depth characteristic of preset threshold, with
Generate pedestrian track figure.
2. pedestrian track generation method as described in claim 1, which is characterized in that described to obtain what each camera detection arrived
The step of pedestrian track information includes:
Obtain the collected tracking frame sequence of each camera；
Determine the sampling tracking frame in each tracking frame sequence；
Pedestrian's template library is updated according to each sampling tracking frame；
Pedestrian track frame is generated according to updated pedestrian's template library；
Obtain the newest pedestrian tracking template in the pedestrian track frame；
According to the corresponding pedestrian track information of each newest pedestrian tracking template generation.
3. pedestrian track generation method as claimed in claim 2, which is characterized in that it is described according to each newest pedestrian with
The step of track template generation corresponding pedestrian track information includes:
The pedestrian image store path of each newest pedestrian tracking template is obtained, and is stored according to each pedestrian image
Path obtains corresponding newest pedestrian image；
The tracking frame in the tracking frame sequence in addition to the sampling tracking frame is carried out according to each newest pedestrian image
Filtering processing, and generate the corresponding pedestrian track information of each newest pedestrian image.
4. pedestrian track generation method as claimed in claim 2, which is characterized in that described according to each sampling tracking frame
Update pedestrian's template library the step of include:
Obtain multiple first pedestrian images in each sampling tracking frame；
Obtain corresponding second pedestrian image of all pedestrian tracking templates in pedestrian's template library；
Obtain the characteristic information of each first pedestrian image and the characteristic information of each second pedestrian image；
Determine that the characteristic information of each first pedestrian image is similar to the characteristic information of each second pedestrian image
Degree；
Pedestrian's template library is updated according to each similarity.
5. pedestrian track generation method as claimed in claim 4, which is characterized in that described to be updated according to each similarity
The step of pedestrian's template library includes:
The similarity of first pedestrian image and multiple second pedestrian images is judged whether there is more than or equal to default
Threshold value；
Determining to be greater than or equal to default threshold there are the similarity of second pedestrian image and multiple first pedestrian images
When value, highest second pedestrian image of similarity is obtained；
Obtain pedestrian's mark in the corresponding pedestrian tracking template of highest second pedestrian image of the similarity；
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to first pedestrian
Pedestrian's template library described in the corresponding pedestrian tracking template renewal of image.
6. pedestrian track generation method as claimed in claim 5, which is characterized in that described to judge whether there is the first row
The similarity of people's image and multiple second pedestrian images was greater than or equal to after the step of preset threshold, further includes:
It is default determining to be greater than or equal to there is no the similarity of first pedestrian image and multiple second pedestrian images
When threshold value, pedestrian's mark is generated by preset algorithm；
It is identified according to the pedestrian and generates the corresponding pedestrian tracking template of first pedestrian image, and according to first pedestrian
Pedestrian's template library described in the corresponding pedestrian tracking template renewal of image.
7. such as the described in any item pedestrian track generation methods of claim 5 or 6, which is characterized in that described according to the pedestrian
Mark generates the step of first pedestrian image corresponding pedestrian tracking template and includes:
Obtain the characteristic information of first pedestrian image；
First pedestrian image is obtained in the location information sampled in tracking frame, the storage road of first pedestrian image
Diameter, the camera identification for generating the sampling tracking frame；
According to pedestrian mark, generate the camera identification for sampling tracking frame, the generation time for sampling tracking frame, right
Location information, the characteristic information of the first pedestrian image and the storage road of first pedestrian image in sampling tracking frame answered
Diameter generates the pedestrian tracking template of first pedestrian image, wherein the generation time of the sampling tracking frame is the camera shooting
Head collects the time obtained when tracking frame sequence.
8. pedestrian track generation method as claimed in claim 7, which is characterized in that generation first pedestrian image
After the step of pedestrian tracking template, further includes:
It is newest pedestrian tracking template by the corresponding pedestrian tracking template-setup of first pedestrian image.
9. pedestrian track generation method as claimed in claim 8, which is characterized in that described that first pedestrian image is corresponding
Pedestrian tracking template-setup be newest pedestrian tracking template the step of after, further includes:
It obtains in each second pedestrian image and is greater than or equal to the similarity of the characteristic information of first pedestrian image
Default similarity, and be the pedestrian tracking template of newest pedestrian tracking template；
It will whether be that the value of newest pedestrian tracking template is set as no in the pedestrian tracking template.
10. pedestrian track generation method as claimed in claim 9, which is characterized in that it is described will be in the pedestrian tracking template
It whether is that the value of newest pedestrian tracking template is set as after no step, further includes:
It obtains and generates the pedestrian tracking template of time and current time difference more than or equal to preset threshold in template library；
Delete the pedestrian tracking template that the difference is greater than or equal to preset threshold.
11. a kind of pedestrian track generating means, which is characterized in that the pedestrian track generating means include processor, memory
And be stored in the pedestrian track that can be run on the memory and on the processor and generate program, the pedestrian track generates
The step of the pedestrian track generation method as described in any one of claims 1 to 10 is realized when program is executed by the processor
Suddenly.
12. a kind of readable storage medium storing program for executing, which is characterized in that it is stored with pedestrian track on the readable storage medium storing program for executing and generates program,
The pedestrian track, which generates, realizes that the pedestrian track as described in any one of claims 1 to 10 is raw when program is executed by processor
The step of at method.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910513473.4A CN110245609A (en) | 2019-06-13 | 2019-06-13 | Pedestrian track generation method, device and readable storage medium storing program for executing |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910513473.4A CN110245609A (en) | 2019-06-13 | 2019-06-13 | Pedestrian track generation method, device and readable storage medium storing program for executing |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN110245609A true CN110245609A (en) | 2019-09-17 |
Family
ID=67887031
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910513473.4A Pending CN110245609A (en) | 2019-06-13 | 2019-06-13 | Pedestrian track generation method, device and readable storage medium storing program for executing |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110245609A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111027462A (en) * | 2019-12-06 | 2020-04-17 | 长沙海格北斗信息技术有限公司 | Pedestrian track identification method across multiple cameras |
| CN111553291A (en) * | 2020-04-30 | 2020-08-18 | 北京爱笔科技有限公司 | Pedestrian trajectory generation method, device, equipment and computer storage medium |
| CN111754604A (en) * | 2020-06-23 | 2020-10-09 | 深圳壹账通智能科技有限公司 | Travel track similarity determination method and related equipment |
| CN114312829A (en) * | 2021-12-06 | 2022-04-12 | 广州文远知行科技有限公司 | Pedestrian trajectory prediction method and device, electronic equipment and storage medium |
| CN115761616A (en) * | 2022-10-13 | 2023-03-07 | 深圳市芯存科技有限公司 | Control method and system based on storage space self-adaption |
| CN117495913A (en) * | 2023-12-28 | 2024-02-02 | 中电科新型智慧城市研究院有限公司 | Cross-space-time correlation method and device for night target track |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108629791A (en) * | 2017-03-17 | 2018-10-09 | 北京旷视科技有限公司 | Pedestrian tracting method and device and across camera pedestrian tracting method and device |
| CN109241227A (en) * | 2018-09-03 | 2019-01-18 | 四川佳联众合企业管理咨询有限公司 | Space-time data based on stacking Ensemble Learning Algorithms predicts modeling method |
| CN109271888A (en) * | 2018-08-29 | 2019-01-25 | 汉王科技股份有限公司 | Personal identification method, device, electronic equipment based on gait |
-
2019
- 2019-06-13 CN CN201910513473.4A patent/CN110245609A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108629791A (en) * | 2017-03-17 | 2018-10-09 | 北京旷视科技有限公司 | Pedestrian tracting method and device and across camera pedestrian tracting method and device |
| CN109271888A (en) * | 2018-08-29 | 2019-01-25 | 汉王科技股份有限公司 | Personal identification method, device, electronic equipment based on gait |
| CN109241227A (en) * | 2018-09-03 | 2019-01-18 | 四川佳联众合企业管理咨询有限公司 | Space-time data based on stacking Ensemble Learning Algorithms predicts modeling method |
Non-Patent Citations (3)
| Title |
|---|
| LIN WU 等: "Deep Recurrent Convolutional Networks for Video-based Person Re-identification: An End-to-End Approach", 《ARXIV》 * |
| YUKE LI: "A Deep Spatiotemporal Perspective for Understanding Crowd Behavior", 《IEEE TRANSACTIONS ON MULTIMEDIA》 * |
| 陈宗海: "《系统仿真技术及应用 第9卷》", 31 August 2007, 中国科学技术大学出版社 * |
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111027462A (en) * | 2019-12-06 | 2020-04-17 | 长沙海格北斗信息技术有限公司 | Pedestrian track identification method across multiple cameras |
| CN111553291A (en) * | 2020-04-30 | 2020-08-18 | 北京爱笔科技有限公司 | Pedestrian trajectory generation method, device, equipment and computer storage medium |
| CN111553291B (en) * | 2020-04-30 | 2023-10-17 | 北京爱笔科技有限公司 | Pedestrian track generation method, device, equipment and computer storage medium |
| CN111754604A (en) * | 2020-06-23 | 2020-10-09 | 深圳壹账通智能科技有限公司 | Travel track similarity determination method and related equipment |
| CN114312829A (en) * | 2021-12-06 | 2022-04-12 | 广州文远知行科技有限公司 | Pedestrian trajectory prediction method and device, electronic equipment and storage medium |
| CN114312829B (en) * | 2021-12-06 | 2024-04-23 | 广州文远知行科技有限公司 | Pedestrian track prediction method and device, electronic equipment and storage medium |
| CN115761616A (en) * | 2022-10-13 | 2023-03-07 | 深圳市芯存科技有限公司 | Control method and system based on storage space self-adaption |
| CN115761616B (en) * | 2022-10-13 | 2024-01-26 | 深圳市芯存科技有限公司 | Control method and system based on storage space self-adaption |
| CN117495913A (en) * | 2023-12-28 | 2024-02-02 | 中电科新型智慧城市研究院有限公司 | Cross-space-time correlation method and device for night target track |
| CN117495913B (en) * | 2023-12-28 | 2024-04-30 | 中电科新型智慧城市研究院有限公司 | Cross-space-time correlation method and device for night target track |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110245609A (en) | Pedestrian track generation method, device and readable storage medium storing program for executing | |
| CN109284733B (en) | Shopping guide negative behavior monitoring method based on yolo and multitask convolutional neural network | |
| US11042991B2 (en) | Determining multiple camera positions from multiple videos | |
| CN110210276A (en) | A kind of motion track acquisition methods and its equipment, storage medium, terminal | |
| CN111954886B (en) | System and method for object tracking | |
| EP4174716A1 (en) | Pedestrian tracking method and device, and computer readable storage medium | |
| CN109299658B (en) | Face detection method, face image rendering device and storage medium | |
| CN107341442A (en) | Motion control method, device, computer equipment and service robot | |
| CN109711241B (en) | Object detection method and device and electronic equipment | |
| WO2020171379A1 (en) | Capturing a photo using a mobile device | |
| CN107408119A (en) | Image retrieving apparatus, system and method | |
| CN112712051B (en) | Object tracking method, device, computer equipment and storage medium | |
| CN112115894A (en) | Training method and device for hand key point detection model and electronic equipment | |
| CN115049731B (en) | Visual image construction and positioning method based on binocular camera | |
| CN114581998A (en) | Deployment and control method, system, equipment and medium based on target object association feature fusion | |
| CN112949539A (en) | Pedestrian re-identification interactive retrieval method and system based on camera position | |
| WO2022033306A1 (en) | Target tracking method and apparatus | |
| CN111798367A (en) | Image processing method, image processing device, storage medium and electronic equipment | |
| JP2009123150A (en) | Object detection apparatus and method, object detection system and program | |
| CN112686195A (en) | Emotion recognition method and device, computer equipment and storage medium | |
| US20230131717A1 (en) | Search processing device, search processing method, and computer program product | |
| CN116580054B (en) | Video data processing method, device, equipment and medium | |
| CN115565253A (en) | Dynamic gesture real-time recognition method and device, electronic equipment and storage medium | |
| CN109886996A (en) | A kind of visual pursuit optimization method | |
| CN115393755A (en) | Visual target tracking method, device, equipment and storage medium |
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: 20190917 |
|
| RJ01 | Rejection of invention patent application after publication |