CN110532883A - On-line tracking is improved using off-line tracking algorithm - Google Patents
On-line tracking is improved using off-line tracking algorithm Download PDFInfo
- Publication number
- CN110532883A CN110532883A CN201910695584.1A CN201910695584A CN110532883A CN 110532883 A CN110532883 A CN 110532883A CN 201910695584 A CN201910695584 A CN 201910695584A CN 110532883 A CN110532883 A CN 110532883A
- Authority
- CN
- China
- Prior art keywords
- tracking
- online
- model
- offline model
- tracked
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/254—Analysis of motion involving subtraction of images
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/40—Analysis of texture
- G06T7/41—Analysis of texture based on statistical description of texture
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/90—Determination of colour characteristics
-
- 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
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V2201/00—Indexing scheme relating to image or video recognition or understanding
- G06V2201/07—Target detection
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02T—CLIMATE CHANGE MITIGATION TECHNOLOGIES RELATED TO TRANSPORTATION
- Y02T10/00—Road transport of goods or passengers
- Y02T10/10—Internal combustion engine [ICE] based vehicles
- Y02T10/40—Engine management systems
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Data Mining & Analysis (AREA)
- Multimedia (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Probability & Statistics with Applications (AREA)
- Human Computer Interaction (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a kind of online tracking, device and computer readable storage medium based on offline model, this method comprises: acquiring Online Video in real time, obtain each frame image in Online Video comprising target to be tracked, and each frame image got is pre-processed, to generate the initial results of target to be tracked, processing is modified to the initial results by preparatory trained off-line tracking model, to generate final result.On the one hand this case is realized before forward trace modification as a result, to improve the precision for handling real-time Online Video stream by combining offline model tracking and in line style tracking by offline model tracking；On the other hand by the real-time Online Video stream of line style tracking processing.
Description
Technical field
The present invention relates to field of computer technology more particularly to a kind of online tracking based on offline model, device and
Computer readable storage medium.
Background technique
Target following and a kind of visual target tracking are an important research directions of current field of machine vision.
Generally, target following detects the interesting target in image sequence, to extract, identify and track, to obtain
The motion state parameters (such as position, speed, acceleration and motion profile etc.) of target to be tracked are taken, so as to further
Processing and analysis are other technologies field (such as vision guided navigation, pose estimation to realize the behavior understanding to moving target
With motion analysis etc.) reference data is provided, target following has extensively in fields such as intelligent monitoring, human-computer interaction, robot navigations
General application.In such applications, target following is robot perception external environment and the basis reacted, is to understand image
Key.
Current goal tracking technique is broadly divided into offline model and in line style two major classes.Wherein offline model is also referred to as batch processing
Type, the central idea of offline model method are exactly the testing result of object in each frame to be connected into small tracking segment, then again
The merging of segment is carried out with relatively reliable feature, more representational offline model method mainly has minimum cost network stream
Algorithm, energy minimization method, and the minimum off-line algorithm of nomography etc. completely；It has been used more just because of offline model method
Information in more before and after frames, can recall before modification as a result, to being finally obtained preferably precision, but also be limited
In this, offline model method can only be handled offline video, cannot handle real-time video flowing.And it is got down in line style method
The matching of target in present frame and next frame just needs immediately to provide when present frame occurs as a result, both can handle in real time
Video flowing also can handle offline video, and line style method can accomplish relatively good real-time, to have in practical applications
One seat；More traditional at present to apply Kalman filtering mostly in line style method, particle filter or Markov are determined
Plan process；But it is often lower than offline model method in the precision of line style method.
Summary of the invention
The present invention provides a kind of online tracking, device and computer readable storage medium based on offline model, master
The precision for being to be promoted the real-time Online Video stream of processing of syllabus.
To achieve the above object, the present invention provides a kind of online tracking based on offline model, this method comprises:
Step A: Online Video is acquired in real time；
Step B: each frame image for containing target to be tracked is obtained from Online Video；
Step C: each frame image got is pre-processed, to generate the initial results of target to be tracked；
Step D: processing is modified to the initial results by preparatory trained off-line tracking model；And
Step E: the final result of target to be tracked is generated.
Optionally, the step C includes:
Step C1: video capture device parameter is pre-seted；
Step C2: the environmental parameter of current scene is acquired；Wherein, the environmental parameter may include, but be not limited to include:
Illumination, tone, noise etc.；
Step C3: it is based on the video capture device parameter and the current scene environmental parameter, for each frame
Image carries out denoising and normalized；
Step C4: the initial results of target to be tracked are generated.
Optionally, the step C further include:
Step C5: to it is described by denoising and normalized each frame image be arranged tracing area, wherein it is described with
Track region is polygon, and the tracing area is the detection zone for including target to be tracked.
Optionally, the environmental parameter includes illumination, tone or noise.
Optionally, the step D includes:
Step D1: the initial results are inputted in trained off-line tracking model in advance.
Optionally, the step D further include:
Step D2: judge whether the initial results need to correct；If it is determined that the initial results need to correct, then hold
Row step D3；
Step D3: the initial results are modified by preparatory trained off-line tracking model.
Optionally, before executing the step D, the online tracking based on offline model track algorithm is also wrapped
Include: training off-line tracking model in advance, the trained off-line tracking model in advance includes off-line tracking algorithm.
Optionally, the principle of the off-line tracking algorithm are as follows:
By the tracking object in each frame of video as a node；
It merges to obtain the similarity measurement between every two object to be tracked by pedestrian's weight identification model and motion model；
Wherein, two tracking objects of the smaller expression of similarity magnitude are more similar.
To achieve the above object, the present invention also provides a kind of based on offline model in sightline tracking device, and described device includes
Memory and processor are stored with the online tracking journey based on offline model that can be run on the processor on the memory
Sequence, when the online trace routine based on offline model is executed by the processor realize as described above based on offline model
Line tracking.
In addition, to achieve the above object, it is described computer-readable the present invention also provides a kind of computer readable storage medium
The online trace routine based on offline model is stored on storage medium, the online trace routine based on offline model can be by one
Or multiple processors execute, the step of to realize online tracking based on offline model as described above.
In online tracking device and computer readable storage medium proposed by the present invention based on offline model, first in fact
When acquire Online Video, then obtain include in Online Video target to be tracked each frame image, and it is each to what is got
Frame image is pre-processed, and to generate the initial results of target to be tracked, then passes through preparatory trained off-line tracking model pair
The initial results are modified processing, to generate final result.This case by combine offline model tracking and line style with
Track method, it is before on the one hand realizing forward trace modification by offline model tracking as a result, real-time to improve processing
The precision of Online Video stream；On the other hand by the real-time Online Video stream of line style tracking processing.
Detailed description of the invention
Fig. 1 is the flow diagram for the online tracking based on offline model that one embodiment of the invention provides；
Fig. 2 is the flow diagram of the step C in Fig. 1；
Fig. 3 is the flow diagram of the step D in Fig. 1；
Fig. 4 is the internal structure chart in sightline tracking device based on offline model that one embodiment of the invention provides；
Fig. 5 is the module diagram for the online trace routine based on offline model that one embodiment of the invention provides.
The object of the invention is realized, the embodiments will be further described with reference to the accompanying drawings for functional characteristics and advantage.
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.
The present invention provides a kind of online tracking based on offline model.Shown in referring to Fig.1, mentioned for one embodiment of the invention
The flow diagram of the online tracking based on offline model supplied.This method can be executed by device, which can be with
By software and or hardware realization.
In the present embodiment, the online tracking based on offline model includes:
Step A: Online Video is acquired in real time；
Step B: each frame image for containing target to be tracked is obtained from Online Video；
Step C: each frame image got is pre-processed, to generate the initial results of target to be tracked；
Step D: processing is modified to the initial results by preparatory trained off-line tracking model；And
Step E: the final result of target to be tracked is generated.
Further, the source of the Online Video in the step A is the video image of video capture device acquisition.
Further, the pretreatment in the step C may include, but be not limited to include: in conjunction with pre-set video
The environmental parameter for acquiring device parameter and current scene, by denoise, normalize etc. processing methods eliminate illumination in different scenes,
The influence of environment tone, noise etc..
Referring to Fig. 2, in more detail, the step C includes:
Step C1: video capture device parameter is pre-seted；
Step C2: the environmental parameter of current scene is acquired；Wherein, the environmental parameter may include, but be not limited to include:
Illumination, tone, noise etc.；
Step C3: it is based on the video capture device parameter and the current scene environmental parameter, for each frame
Image carries out denoising and normalized；
Step C4: the initial results of target to be tracked are generated.
Further, the step C further include:
Step C5: tracing area is set by each frame image of denoising and normalized to described.Wherein, it is described with
Track region can be the polygon of arbitrary shape, and the tracing area is the detection zone for including target to be tracked.
Further, the target to be tracked in the step C can be personage, animal, plant, object etc..The personage
May be, but not limited to, is pedestrian, the people of working condition, people by bus, the people of driving or the people on the default vehicles
Deng.The animal may be, but not limited to, the animals such as cat, dog, pig, bird, fish.The plant may be, but not limited to, be flower,
The objects such as grass, trees.It is the object with certain form such as computer, barcode scanning equipment, balloon that the object, which may be, but not limited to,
Body.
In the present embodiment, the target to be tracked is illustrated by taking pedestrian as an example.
Further, before executing the step D, the online tracking based on offline model track algorithm is also wrapped
It includes: training off-line tracking model in advance.The trained off-line tracking model in advance includes off-line tracking algorithm.
Specifically, the principle of the off-line tracking algorithm are as follows:
By the tracking object in each frame of video as a node, (Person Re- is then identified by pedestrian again
Identification, referred to as ReID) model and motion model merge to obtain the similitude between every two object to be tracked
Measurement.Wherein, two tracking objects of the smaller expression of similarity magnitude are more similar.
About two concepts of current object detection field, hand over and than (Intersection-over-Union, IoU) and
Selective search (Selective Search), briefly describes:
One, it hands over and compares, be the candidate frame (candidate bound) generated and former indicia framing (ground truth
Bound overlapping rate), the i.e. ratio of their intersection and union；Most ideally completely overlapped, i.e., ratio is 1.
Two, in selective search (Selective Search), there are three types of strategies, are respectively:
By using the various colors space with different invariancies；
By using different similarity measurements；And
By using different initialization areas.
Wherein, in one embodiment, about similarity measurement, there is different combinations, such as there are following four kinds of measurements
Algorithm:
1) color similarity: S (color), wherein the metric algorithm of S (color) expression color similarity；
2) texture similarity: S (texture), wherein the metric algorithm of S (texture) expression color similarity；
3) size similarity: S (size), wherein the metric algorithm of S (size) expression color similarity；
4) identical similarity: S (fit), wherein the metric algorithm of S (fit) expression color similarity；
Above-mentioned four kinds of metric algorithms are merged into a kind of strategy: S=a*S (color)+b*S by certain mode
(texture)+c*S(size)+d*S(fit)。
In practical applications, since the object to be tracked in same frame can not be associated with, also, if one to be tracked right
As A and object B to be tracked are same targets, while object A to be tracked and object C to be tracked are same targets, then to be tracked
Object B is also same target with object C to be tracked.Under these constraints, tracking problem a two-value planning is converted to ask
Topic.In the present embodiment, two-value planning problem is solved using Gurob.Gurob is a kind of scale mathematical plan optimization device.Institute
Two-value planning application is stated in frame differential method, since the target in scene is moving, the image of target is in different images frame
Position it is different.Frame differential method carries out calculus of differences, the corresponding picture of different frame to time upper continuous two frame or three frame images
Vegetarian refreshments subtracts each other, and judges gray scale absolute value of the difference, when absolute value is more than certain threshold value, can be judged as moving target, thus real
The detection function of existing target.Its concrete principle are as follows: remember n-th frame and the (n-1)th frame image is respectively fn and fn-1, two frame respective pixels
The gray value of point is denoted as fn (x, y) and fn-1 (x, y) respectively, according to following formula by the gray value of two field pictures corresponding pixel points
Subtracted each other, and take its absolute value, obtains difference image Dn:
Wherein, Dn (x, y)=| fn (x, y)-fn-1 (x, y) |
Given threshold T carries out binary conversion treatment to pixel one by one according to following formula, obtains binary image Rn '.Its
In, the point that gray value is 255 is prospect (target to be tracked) point, and the point that gray value is 0 is background dot；To image Rn ' into
Row connectivity analysis can finally obtain the image Rn containing entire motion target.
As Dn (x, y) > T, Rn ' (x, y)=255；
Otherwise, Rn ' (x, y)=0.
Further, the step D includes:
Step D1: the initial results are inputted in trained off-line tracking model in advance；
Step D2: judge whether the initial results need to correct；If it is determined that the initial results need to correct, then hold
Row step D3；If it is determined that the initial results do not need to correct, D4 is thened follow the steps.
Step D3: the initial results are modified by preparatory trained off-line tracking model；
Step D4: the initial results are not modified.
In this embodiment, it after being modified by preparatory trained off-line tracking model to the initial results, produces
The final result of raw target to be tracked.
Specifically, in one embodiment, the principle of the step D are as follows: (consider calculation amount and reality every preset quantity frame
Shi Xing, not to every frame, such as every four frames) just to the initial results generated, using the off-line tracking model it is offline with
Track algorithm is recalculated, and then carries out continuing tracking calculating in current results, to obtain final result.More specifically,
When the current similarity magnitude recalculated using off-line tracking algorithm is smaller than upper similarity magnitude, then institute is judged
Initial results are stated to need to correct.It is to be understood that in the present embodiment, the off-line tracking algorithm be it is in the prior art from
Line track algorithm, the embodiment of the present invention are not specifically limited off-line tracking algorithm.
Further, in the present embodiment, it is illustrated by taking pedestrian as an example, the mesh to be tracked generated in the step E
Target final result includes position and the quantity of pedestrian.
Online tracking provided by the invention based on offline model acquires Online Video in real time first, then obtains
Include each frame image of target to be tracked in line video, and each frame image got pre-processed, with generate to
The initial results of target are tracked, then processing is modified to the initial results by preparatory trained off-line tracking model,
To generate final result.This case by combining offline model tracking and in line style tracking, on the one hand by offline model with
Track method is realized before forward trace modification as a result, to improve the precision for handling real-time Online Video stream；On the other hand
By in the real-time Online Video stream of line style tracking processing.
The present invention also provides a kind of based on offline model in sightline tracking device.It is one embodiment of the invention referring to shown in Fig. 4
The internal structure chart in sightline tracking device based on offline model provided.Can be in sightline tracking device based on offline model
PC (Personal Computer, PC), is also possible to the terminals such as smart phone, tablet computer, portable computer and sets
It is standby.The code library managing device includes at least memory 11, processor 12, network interface 13 and communication bus 14.
Wherein, the memory 11 includes at least a type of readable storage medium storing program for executing, and the readable storage medium storing program for executing includes
Flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory etc.), magnetic storage, disk, CD etc..Institute
It states memory 11 in some embodiments and can be the internal storage unit in sightline tracking device based on offline model, such as the base
In the hard disk in sightline tracking device of offline model.The memory 11 is also possible in further embodiments based on offline model
The plug-in type hard disk being equipped on sightline tracking device in the External memory equipment of sightline tracking device, such as based on offline model, intelligence
Energy storage card (Smart Media Card, SMC), secure digital (Secure Digital, SD) card, flash card (Flash
Card) etc..Further, the memory 11 can also both include the storage inside list in sightline tracking device based on offline model
Member also includes External memory equipment.The memory 11 can be not only used for storage and be installed on the online tracking dress based on offline model
The application software and Various types of data set, such as the code etc. of the online trace routine based on offline model, can be also used for temporarily
Store the data that has exported or will export.
The processor 12 can be in some embodiments a central processing unit (Central Processing Unit,
CPU), controller, microcontroller, microprocessor or other data processing chips store in the memory 11 for running
Program code or processing data, such as execute the online trace routine etc. based on offline model.
The network interface 13 optionally may include standard wireline interface and wireless interface (such as WI-FI interface), usually
For establishing communication connection between the marketing clue extraction element and other electronic equipments that understand based on reading.
The communication bus 14 is for realizing the connection communication between these components.
Fig. 4 illustrate only the online trace routine with component 11 to 14 and based on offline model based on offline model
Sightline tracking device, it will be appreciated by persons skilled in the art that the structure shown in Fig. 4 is not constituted to based on the online of offline model
The restriction of tracking device may include perhaps combining certain components or different portions than illustrating less perhaps more components
Part arrangement.
It is shown in Fig. 4 based on offline model in sightline tracking device embodiment, be stored with and be based in the memory 11
The online trace routine of offline model；The processor 12 executes the online tracking based on offline model stored in the memory 11
Following steps are realized when program:
Step A: Online Video is acquired in real time；
Step B: each frame image for containing target to be tracked is obtained from Online Video；
Step C: each frame image got is pre-processed, to generate the initial results of target to be tracked；
Step D: processing is modified to the initial results by preparatory trained off-line tracking model；And
Step E: the final result of target to be tracked is generated.
The online trace routine based on offline model can be divided into one or more functions according to its different function
Module.One or more module is stored in the memory 11, and by one or more processors (the present embodiment be
Reason device 12) it is performed to complete the present invention, the so-called module of the present invention is the series of computation machine for referring to complete specific function
Program instruction section, for describing the online trace routine based on offline model in the execution in sightline tracking device based on offline model
Process.
For example, referring to shown in Fig. 5, for the present invention is based on offline models in one embodiment of sightline tracking device based on offline
The program module schematic diagram of the online trace routine of type, in the embodiment, the online trace routine based on offline model can be divided
It is segmented into video acquisition module 31, frame image collection module 32, preprocessing module 33, offline tracing module 34 and result generation module
35, illustratively:
Video acquisition module 31, for acquiring Online Video in real time；
Frame image collection module 32, for obtaining each frame image for containing target to be tracked from Online Video；
Preprocessing module 33, for being pre-processed to each frame image got, to generate the first of target to be tracked
Beginning result；
Offline tracing module 34, for being modified by preparatory trained off-line tracking model to the initial results
Processing；
As a result generation module 35, for generating the final result of target to be tracked.
Above-mentioned video acquisition module 31, frame image collection module 32, preprocessing module 33, offline tracing module 34 and result
The program modules such as generation module 35 are performed realized functions or operations step and are substantially the same with above-described embodiment, herein not
It repeats again.
Fig. 5 illustrate only the online trace routine with module 31-35 and based on offline model based on offline model
Sightline tracking device, it will be appreciated by persons skilled in the art that the structure shown in Fig. 5 was not constituted to described based on offline model
It may include perhaps combining certain module or difference than illustrating less perhaps more modules in the restriction of sightline tracking device
Module arrangement.
It, can also be in addition, each functional module in each embodiment of the present invention can integrate in one processing unit
It is that each unit physically exists alone, can also be integrated in one unit with two or more units.It is above-mentioned it is integrated can
To use formal implementation of hardware, can also be realized in the form of hardware adds software function module.
In addition, the embodiment of the present invention also proposes a kind of computer readable storage medium, the computer readable storage medium
On be stored with the online trace routine based on offline model, the online trace routine based on offline model can be by one or more
Device is managed to execute, to realize following operation:
Step A: Online Video is acquired in real time；
Step B: each frame image for containing target to be tracked is obtained from Online Video；
Step C: each frame image got is pre-processed, to generate the initial results of target to be tracked；
Step D: processing is modified to the initial results by preparatory trained off-line tracking model；And
Step E: the final result of target to be tracked is generated.
Computer readable storage medium specific embodiment of the present invention and it is above-mentioned based on offline model in sightline tracking device and
Each embodiment of method is essentially identical, does not make tired state herein.
It should be noted that the serial number of the above embodiments of the invention is only for description, do not represent the advantages or disadvantages of the embodiments.And
Term " includes " herein or any other variant thereof is intended to cover non-exclusive inclusion, so that including a series of
Process, device, article or the method for element not only include those elements, but also other including being not explicitly listed are wanted
Element, or further include for this process, device, article or the intrinsic element of method.The case where not limiting more
Under, the element that is limited by sentence " including ... ", it is not excluded that in process, device, article or the method for including the element
There is also other identical elements.
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 terminal device (it can be mobile phone,
Computer, server 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 (10)
1. a kind of online tracking based on offline model, which is characterized in that the online tracking packet based on offline model
It includes:
Step A: Online Video is acquired in real time；
Step B: each frame image for containing target to be tracked is obtained from Online Video；
Step C: each frame image got is pre-processed, to generate the initial results of target to be tracked；
Step D: processing is modified to the initial results by preparatory trained off-line tracking model；And
Step E: the final result of target to be tracked is generated.
2. the online tracking based on offline model as described in claim 1, which is characterized in that the step C includes:
Step C1: video capture device parameter is pre-seted；
Step C2: the environmental parameter of current scene is acquired；Wherein, the environmental parameter may include, but be not limited to include: light
According to, tone, noise etc.；
Step C3: it is based on the video capture device parameter and the current scene environmental parameter, for each frame image
Carry out denoising and normalized；
Step C4: the initial results of target to be tracked are generated.
3. the online tracking based on offline model as claimed in claim 2, which is characterized in that the step C further include:
Step C5: tracing area is set by each frame image of denoising and normalized to described, wherein the tracking area
Domain is polygon, and the tracing area is the detection zone for including target to be tracked.
4. the online tracking based on offline model as claimed in claim 2 or claim 3, which is characterized in that the environmental parameter packet
Include illumination, tone or noise.
5. the online tracking as claimed in claim 1 or 3 based on offline model, which is characterized in that the step D includes:
Step D1: the initial results are inputted in trained off-line tracking model in advance.
6. the online tracking based on offline model as claimed in claim 5, which is characterized in that the step D further include:
Step D2: judge whether the initial results need to correct；If it is determined that the initial results need to correct, then step is executed
Rapid D3；
Step D3: the initial results are modified by preparatory trained off-line tracking model.
7. the online tracking based on offline model as described in claim 1, which is characterized in that execute the step D it
Before, the online tracking based on offline model track algorithm further include: training off-line tracking model in advance, the preparatory instruction
The off-line tracking model perfected includes off-line tracking algorithm.
8. the online tracking based on offline model as claimed in claim 7, which is characterized in that the off-line tracking algorithm
Principle are as follows:
By the tracking object in each frame of video as a node；
It merges to obtain the similarity measurement between every two object to be tracked by pedestrian's weight identification model and motion model；
Wherein, two tracking objects of the smaller expression of similarity magnitude are more similar.
9. it is a kind of based on offline model in sightline tracking device, which is characterized in that it is described based on offline model in sightline tracking device packet
Memory and processor are included, the online tracking based on offline model that can be run on the processor is stored on the memory
Program is realized as described in claim 1-8 is any when the online trace routine based on offline model is executed by the processor
The online tracking based on offline model.
10. a kind of computer readable storage medium, which is characterized in that be stored on the computer readable storage medium based on from
The online trace routine of line style, the online trace routine based on offline model can be executed by one or more processor, with
The step of realizing the online tracking as claimed in any one of claims 1 to 8 based on offline model.
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910695584.1A CN110532883B (en) | 2019-07-30 | 2019-07-30 | Improvement of on-line tracking algorithm by off-line tracking algorithm |
| PCT/CN2019/117538 WO2021017283A1 (en) | 2019-07-30 | 2019-11-12 | Offline method-based online tracking method and apparatus, computer device, and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910695584.1A CN110532883B (en) | 2019-07-30 | 2019-07-30 | Improvement of on-line tracking algorithm by off-line tracking algorithm |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110532883A true CN110532883A (en) | 2019-12-03 |
| CN110532883B CN110532883B (en) | 2023-09-01 |
Family
ID=68661109
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910695584.1A Active CN110532883B (en) | 2019-07-30 | 2019-07-30 | Improvement of on-line tracking algorithm by off-line tracking algorithm |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110532883B (en) |
| WO (1) | WO2021017283A1 (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814725A (en) * | 2020-07-20 | 2020-10-23 | 北京华正明天信息技术股份有限公司 | Early warning method for judging ignition of monitoring video based on CNN + LSTM + MLP combined neural network |
| CN112131966A (en) * | 2020-09-01 | 2020-12-25 | 深圳中兴网信科技有限公司 | Mud truck monitoring method and system and storage medium |
| CN112612768A (en) * | 2020-12-11 | 2021-04-06 | 上海哔哩哔哩科技有限公司 | Model training method and device |
| CN113658210A (en) * | 2021-09-02 | 2021-11-16 | 西安中科西光航天科技有限公司 | Front-end real-time target tracking method based on Jetson NX platform |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112861971A (en) * | 2021-02-07 | 2021-05-28 | 启迪云控(上海)汽车科技有限公司 | Cross-point road side perception target tracking method and system |
| CN113744299B (en) * | 2021-09-02 | 2022-07-12 | 上海安维尔信息科技股份有限公司 | Camera control method and device, electronic equipment and storage medium |
| CN114092516B (en) * | 2021-11-08 | 2024-05-14 | 国汽智控(北京)科技有限公司 | Multi-target tracking detection method, device, equipment and medium |
| CN114495612B (en) * | 2021-12-15 | 2023-12-26 | 华中光电技术研究所（中国船舶重工集团公司第七一七研究所） | Online simulation training device for infrared tracking warning equipment |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101325691A (en) * | 2007-06-14 | 2008-12-17 | 清华大学 | Method and apparatus for tracing a plurality of observation model with fusion of differ durations |
| CN104134078A (en) * | 2014-07-22 | 2014-11-05 | 华中科技大学 | Automatic selection method for classifiers in people flow counting system |
| US20150371102A1 (en) * | 2014-06-18 | 2015-12-24 | Delta Electronics, Inc. | Method for recognizing and locating object |
| CN107679455A (en) * | 2017-08-29 | 2018-02-09 | 平安科技（深圳）有限公司 | Target tracker, method and computer-readable recording medium |
| CN109447121A (en) * | 2018-09-27 | 2019-03-08 | 清华大学 | A kind of Visual Sensor Networks multi-object tracking method, apparatus and system |
| CN109636829A (en) * | 2018-11-24 | 2019-04-16 | 华中科技大学 | A kind of multi-object tracking method based on semantic information and scene information |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103218816B (en) * | 2013-04-18 | 2016-05-04 | 中山大学 | A kind of crowd density estimation method and people flow rate statistical method based on video analysis |
| CN109800624A (en) * | 2018-11-27 | 2019-05-24 | 上海眼控科技股份有限公司 | A kind of multi-object tracking method identified again based on pedestrian |
-
2019
- 2019-07-30 CN CN201910695584.1A patent/CN110532883B/en active Active
- 2019-11-12 WO PCT/CN2019/117538 patent/WO2021017283A1/en active Application Filing
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101325691A (en) * | 2007-06-14 | 2008-12-17 | 清华大学 | Method and apparatus for tracing a plurality of observation model with fusion of differ durations |
| WO2008151577A1 (en) * | 2007-06-14 | 2008-12-18 | Tsinghua University | Tracking method and device adopting a series of observation models with different lifespans |
| US20150371102A1 (en) * | 2014-06-18 | 2015-12-24 | Delta Electronics, Inc. | Method for recognizing and locating object |
| CN104134078A (en) * | 2014-07-22 | 2014-11-05 | 华中科技大学 | Automatic selection method for classifiers in people flow counting system |
| CN107679455A (en) * | 2017-08-29 | 2018-02-09 | 平安科技（深圳）有限公司 | Target tracker, method and computer-readable recording medium |
| WO2019041519A1 (en) * | 2017-08-29 | 2019-03-07 | 平安科技（深圳）有限公司 | Target tracking device and method, and computer-readable storage medium |
| CN109447121A (en) * | 2018-09-27 | 2019-03-08 | 清华大学 | A kind of Visual Sensor Networks multi-object tracking method, apparatus and system |
| CN109636829A (en) * | 2018-11-24 | 2019-04-16 | 华中科技大学 | A kind of multi-object tracking method based on semantic information and scene information |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111814725A (en) * | 2020-07-20 | 2020-10-23 | 北京华正明天信息技术股份有限公司 | Early warning method for judging ignition of monitoring video based on CNN + LSTM + MLP combined neural network |
| CN112131966A (en) * | 2020-09-01 | 2020-12-25 | 深圳中兴网信科技有限公司 | Mud truck monitoring method and system and storage medium |
| CN112612768A (en) * | 2020-12-11 | 2021-04-06 | 上海哔哩哔哩科技有限公司 | Model training method and device |
| CN113658210A (en) * | 2021-09-02 | 2021-11-16 | 西安中科西光航天科技有限公司 | Front-end real-time target tracking method based on Jetson NX platform |
Also Published As
| Publication number | Publication date |
|---|---|
| WO2021017283A1 (en) | 2021-02-04 |
| CN110532883B (en) | 2023-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110532883A (en) | On-line tracking is improved using off-line tracking algorithm | |
| Biggs et al. | Who left the dogs out? 3d animal reconstruction with expectation maximization in the loop | |
| CN106709461B (en) | Activity recognition method and device based on video | |
| CN106941602B (en) | Locomotive driver behavior identification method and device | |
| CN108197589B (en) | Semantic understanding method, apparatus, equipment and the storage medium of dynamic human body posture | |
| Hernández-Vela et al. | Graph cuts optimization for multi-limb human segmentation in depth maps | |
| CN110751022A (en) | Urban pet activity track monitoring method based on image recognition and related equipment | |
| CN106648078B (en) | Multi-mode interaction method and system applied to intelligent robot | |
| CN111626126A (en) | Face emotion recognition method, device, medium and electronic equipment | |
| CN109241890B (en) | Face image correction method, apparatus and storage medium | |
| CN109670517A (en) | Object detection method, device, electronic equipment and target detection model | |
| CN110688897A (en) | Pedestrian re-identification method and device based on joint judgment and generation learning | |
| CN111062328A (en) | Image processing method and device and intelligent robot | |
| CN108073851A (en) | A kind of method, apparatus and electronic equipment for capturing gesture identification | |
| CN112668492A (en) | Behavior identification method for self-supervised learning and skeletal information | |
| CN110533699A (en) | The dynamic multiframe speed-measuring method of pixel variation based on optical flow method | |
| CN114064974A (en) | Information processing method, information processing apparatus, electronic device, storage medium, and program product | |
| CN112541870A (en) | Video processing method and device, readable storage medium and electronic equipment | |
| Browning et al. | Cortical dynamics of navigation and steering in natural scenes: Motion-based object segmentation, heading, and obstacle avoidance | |
| CN110533688A (en) | Follow-on method for tracking target, device and computer readable storage medium | |
| CN110659576A (en) | Pedestrian searching method and device based on joint judgment and generation learning | |
| CN115862130A (en) | Behavior recognition method based on human body posture and body motion field thereof | |
| Vignolo et al. | The complexity of biological motion | |
| CN114359892A (en) | Three-dimensional target detection method and device and computer readable storage medium | |
| CN114495272A (en) | Motion recognition method, motion recognition device, storage medium, and computer apparatus |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |