CN108122026A - The accurate tracking of attack vehicle holder - Google Patents
The accurate tracking of attack vehicle holder Download PDFInfo
- Publication number
- CN108122026A CN108122026A CN201711369708.4A CN201711369708A CN108122026A CN 108122026 A CN108122026 A CN 108122026A CN 201711369708 A CN201711369708 A CN 201711369708A CN 108122026 A CN108122026 A CN 108122026A
- Authority
- CN
- China
- Prior art keywords
- tracking
- target
- algorithms
- kcf
- dsst
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D3/00—Control of position or direction
- G05D3/12—Control of position or direction using feedback
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/20—Processor architectures; Processor configuration, e.g. pipelining
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Molecular Biology (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Automation & Control Theory (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Image Analysis (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a kind of accurate tracking of attack vehicle holder, for solving the technical issues of existing unmanned vehicle tracking tracking accuracy is poor.Technical solution is to install camera additional on attack vehicle holder, using the classificatory scale space tracker (DSST of correlation filter, discriminative scale space tracker), i.e. by traditional KCF Algorithms Integrations into particle filter frame, the motion state of moving target is using the method predicted in real time, for the KCF algorithms of fixed scale moving target can only be tracked, DSST algorithms are incorporated wherein, carry out size estimation.While accurately being tracked to mutative scale target, the detection of target and the hardware of recognizer are based on SDP and FPGA.The present invention can effectively improve the tracking target capability of attack vehicle, tracking accuracy is good using DSST algorithm improvement KCF algorithms to the object of mutative scale into line trace.
Description
Technical field
The present invention relates to a kind of unmanned vehicle tracking, more particularly to a kind of accurate tracking of attack vehicle holder.
Background technology
Attack vehicle holder is that attack vehicle realizes target following, target identification, target attack and the weight for controlling attack vehicle action
Want device.In the case of small space and Multi-target Attacking, realizing the accurate tracking of target needs advanced algorithm to support, energy
Enough fast and effeciently target followings are the key that attack vehicle performance quality and the groundwork function of attack vehicle holder.
Document " He Meng,《Unmanned vehicle track algorithm research based on machine vision》[D] Beijing Jiaotong University, 2017. " is right
The Visual servoing control algorithm of monopod video camera is studied, and devises a kind of holder movement based on image coordinate offset feedback
Control algolithm.The algorithm makes holder that can rotate tracking target according to target relative position actuated camera, so as to ensure target mark
Target barycenter always in image middle position, efficiently solve the fixed camera coverage of tradition during unmanned vehicle tracking be limited,
The problem of BREAK TRACK.But in practice, attack vehicle holder requirement can be tracked accurately for the mutative scale of object, this method
It can not realize.
The content of the invention
In order to overcome the shortcomings of that existing unmanned vehicle tracking tracking accuracy is poor, it is accurate that the present invention provides a kind of attack vehicle holder
True tracking.This method installs camera additional on attack vehicle holder, using the classificatory scale space tracker of correlation filter
(DSST, discriminative scale space tracker), i.e., by traditional KCF Algorithms Integrations to particle filter frame
In, the motion state of moving target is calculated using the method predicted in real time for the KCF that can only track fixed scale moving target
DSST algorithms are incorporated wherein, carry out size estimation by method.While accurately being tracked to mutative scale target, the inspection of target
It surveys and is based on SDP and FPGA with the hardware of recognizer.The present invention, can be to mutative scale using DSST algorithm improvement KCF algorithms
Object effectively improves the tracking target capability of attack vehicle into line trace, and tracking accuracy is good.
The technical solution adopted by the present invention to solve the technical problems:A kind of accurate tracking of attack vehicle holder, it is special
Point is to comprise the following steps:
The first step, the camera review sent by infrared detector and row, field enable and reset signal initially enters FPGA;
Second step gives dual port RAM after handling camera review;
3rd step, from main DSP after the signal and image that FPGA is sent is received, start to process signal and picture number
According to two panels dual port RAM uses the working method of Pingpang Memory, i.e., when a piece of dual port RAM is receiving the image of preprocessor transmission
During data, another dual port RAM is for tracing figure as tracking module provides the data of image procossing.Tracking module and preprocessing module
Between shaken hands by multichannel buffered serial port, main extension of the image trace module using large capacity SDRAM as main DSP is deposited
Reservoir stores the various intermediate data in image processing process;
4th step includes the algorithm for handling camera review in main DSP, quickly tracks target for KCF algorithms, uses
The practicability wave filter of solution posterior probability based on bayesian theory and monte carlo method, particle filter, to KCF algorithms
It is improved and then to target into line trace, first initializes particle, KCF and DSST parameters, judge the cycling condition of particle every time,
Judge with KCF into line trace, keep particle state, find optimal particle value, used on the basis of optimal particle value
DSST algorithms update target scale, then population, KCF algorithms and DSST algorithms are updated, and judge whether to complete tracking.It follows
The 4th step of ring is until meeting tracer request;
5th step, the execution unit that the result of processing is sent instructions to equipment.
The beneficial effects of the invention are as follows:This method installs camera additional on attack vehicle holder, using point of correlation filter
Class metric space tracker (DSST, discriminative scale space tracker), i.e., it is traditional KCF algorithms is whole
It closes in particle filter frame, the motion state of moving target is using the method predicted in real time, for that can only track fixed ruler
The KCF algorithms of moving target are spent, DSST algorithms are incorporated wherein, carry out size estimation.To mutative scale target carry out accurately with
While track, the detection of target and the hardware of recognizer are based on SDP and FPGA.The present invention is calculated using DSST algorithm improvements KCF
Method can effectively improve the tracking target capability of attack vehicle, tracking accuracy is good to the object of mutative scale into line trace.
It elaborates with reference to the accompanying drawings and detailed description to the present invention.
Description of the drawings
Fig. 1 is the hardware system master-plan block diagram in the accurate tracking of attack vehicle holder of the present invention.
Fig. 2 is KCF the and DSST control algolithm flow charts in the method for the present invention.
Specific embodiment
With reference to Fig. 1-2.The accurate tracking of attack vehicle holder of the present invention is as follows:
Using main DSP as core, two pieces are coprocessor from DSP, complete the identification to target and real-time tracking jointly.Wherein
Preprocessing module is using a piece of main DSP and a piece of FPGA as core processor.
The first step, the camera review and row sent by infrared detector, field enable and reset signal initially enters FPGA;
Second step gives dual port RAM after handling camera review;
3rd step, from signals and image of the main DSP after the signal and image that FPGA sends is received in start to process
Data, the working method that two panels dual port RAM is stored using " table tennis ", i.e., in order to ensure the real-time of image procossing, when a piece of double
For mouth RAM when receiving the image data that preprocessor is sent, another dual port RAM is for tracing figure as tracking module is provided at image
The data of reason.It is shaken hands between tracking module and preprocessing module by multichannel buffered serial port, image trace module uses
Main extended menories of the large capacity SDRAM as main DSP stores the various intermediate data in image processing process;
4th step includes the algorithm for handling camera review in main DSP, target is quickly tracked for KCF algorithms, uses
A kind of practicability wave filter of the solution posterior probability based on bayesian theory and monte carlo method, particle filter, to KCF
Algorithm is improved to target into line trace, first initializes particle, KCF and DSST parameters, judges the cycling condition of particle every time,
Judge with KCF into line trace, keep particle state, find optimal particle value, used on the basis of optimal particle value
DSST algorithms update target scale, then population, KCF algorithms and DSST algorithms are updated, and judge whether to complete tracking.It follows
The ring above process is until meeting tracer request;
The result of processing is sent instructions to the execution unit of equipment by the 5th step.
Claims (1)
1. a kind of accurate tracking of attack vehicle holder, it is characterised in that comprise the following steps:
The first step, the camera review sent by infrared detector and row, field enable and reset signal initially enters FPGA;
Second step gives dual port RAM after handling camera review;
3rd step, from main DSP after the signal and image that FPGA is sent is received, start to process signal and image data, two
Piece dual port RAM uses the working method of Pingpang Memory, i.e., when a piece of dual port RAM is receiving the image data of preprocessor transmission
When, another dual port RAM is for tracing figure as tracking module provides the data of image procossing;Between tracking module and preprocessing module
It is shaken hands by multichannel buffered serial port, image trace module uses main extension storages of the large capacity SDRAM as main DSP
Device stores the various intermediate data in image processing process;
4th step, main DSP include handle camera review algorithm, quickly track target for KCF algorithms, using based on
The practicability wave filter of the solution posterior probability of bayesian theory and monte carlo method, particle filter carry out KCF algorithms
It improves and then to target into line trace, first initializes particle, KCF and DSST parameters, judge the cycling condition of particle every time, use
KCF judges into line trace, keeps particle state, finds optimal particle value, and DSST is used on the basis of optimal particle value
Algorithm updates target scale, then population, KCF algorithms and DSST algorithms are updated, and judges whether to complete tracking;Cycle the
Four steps are until meeting tracer request;
5th step, the execution unit that the result of processing is sent instructions to equipment.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711369708.4A CN108122026A (en) | 2017-12-19 | 2017-12-19 | The accurate tracking of attack vehicle holder |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711369708.4A CN108122026A (en) | 2017-12-19 | 2017-12-19 | The accurate tracking of attack vehicle holder |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108122026A true CN108122026A (en) | 2018-06-05 |
Family
ID=62230343
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711369708.4A Pending CN108122026A (en) | 2017-12-19 | 2017-12-19 | The accurate tracking of attack vehicle holder |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108122026A (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109194935A (en) * | 2018-11-14 | 2019-01-11 | 众格智能科技(上海)有限公司 | A kind of target tracker |
CN109859243A (en) * | 2019-01-18 | 2019-06-07 | 昆明理工大学 | A kind of motion target tracking method based on dimension self-adaption block particle |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101447075A (en) * | 2008-12-31 | 2009-06-03 | 天津理工大学 | Wide-angle lens-based FPGA & DSP embedded multi-valued targets threshold categorization tracking device |
CN203352713U (en) * | 2013-07-22 | 2013-12-18 | 南通大学 | Related tracking device based on DSP |
CN103826105A (en) * | 2014-03-14 | 2014-05-28 | 贵州大学 | Video tracking system and realizing method based on machine vision technology |
-
2017
- 2017-12-19 CN CN201711369708.4A patent/CN108122026A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101447075A (en) * | 2008-12-31 | 2009-06-03 | 天津理工大学 | Wide-angle lens-based FPGA & DSP embedded multi-valued targets threshold categorization tracking device |
CN203352713U (en) * | 2013-07-22 | 2013-12-18 | 南通大学 | Related tracking device based on DSP |
CN103826105A (en) * | 2014-03-14 | 2014-05-28 | 贵州大学 | Video tracking system and realizing method based on machine vision technology |
Non-Patent Citations (3)
Title |
---|
朱航江 等: ""运动状态与尺度估计的核相关目标跟踪方法"", 《计算机科学》 * |
金纯: ""基于无人机平台的目标检测跟踪***研究"", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
钱堂慧 等: ""核相关滤波跟踪算法的尺度自适应改进"", 《计算机应用》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109194935A (en) * | 2018-11-14 | 2019-01-11 | 众格智能科技(上海)有限公司 | A kind of target tracker |
CN109859243A (en) * | 2019-01-18 | 2019-06-07 | 昆明理工大学 | A kind of motion target tracking method based on dimension self-adaption block particle |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20200159256A1 (en) | Method for detecting target object, detection apparatus and robot | |
CN105760824B (en) | A kind of moving human hand tracking method and system | |
CN110849367A (en) | Indoor positioning and navigation method based on visual SLAM fused with UWB | |
CN105913028B (en) | Face + + platform-based face tracking method and device | |
Gao et al. | Firefly algorithm (FA) based particle filter method for visual tracking | |
Mueggler et al. | Towards evasive maneuvers with quadrotors using dynamic vision sensors | |
Li et al. | Vision-aided inertial navigation for resource-constrained systems | |
CN101344965A (en) | Tracking system based on binocular camera shooting | |
CN110139031B (en) | Video anti-shake system based on inertial sensing and working method thereof | |
CN103997624A (en) | Overlapped domain dual-camera target tracking system and method | |
CN108122026A (en) | The accurate tracking of attack vehicle holder | |
CN204305213U (en) | The interactive sighting device of multi-cam human body tracking | |
CN107592508A (en) | A kind of intelligent track identification system | |
Schmidt et al. | An indoor RGB-D dataset for the evaluation of robot navigation algorithms | |
CN102214301A (en) | Multi-target tracking method for associated cooperation of adaptive motion | |
CN108053422A (en) | Mobile target monitoring method | |
Yang et al. | Delivery of pollen to forsythia flower pistils autonomously and precisely using a robot arm | |
Liu et al. | Accurate real-time ball trajectory estimation with onboard stereo camera system for humanoid ping-pong robot | |
CN110377033B (en) | RGBD information-based small football robot identification and tracking grabbing method | |
CN106228571B (en) | The object tracking detection method and device of object manipulator | |
TWI520110B (en) | 3d visual detection system and method for determining if an object enters a zone on demand | |
Wu et al. | Interactive multi-camera soccer video analysis system | |
CN100362531C (en) | Real-time automatic moving portrait tracking method incorporating time domain differential and spatial domain diversity | |
CN110493618A (en) | Android method for intelligently controlling televisions and system based on USB3.0 interface | |
CN109785365A (en) | Address events drive the real-time modeling method method of unstructured signal |
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 | ||
WD01 | Invention patent application deemed withdrawn after publication | ||
WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180605 |