CN108985233A - One kind being based on the relevant high-precision wireless vehicle tracking of digital picture - Google Patents
One kind being based on the relevant high-precision wireless vehicle tracking of digital picture Download PDFInfo
- Publication number
- CN108985233A CN108985233A CN201810795696.XA CN201810795696A CN108985233A CN 108985233 A CN108985233 A CN 108985233A CN 201810795696 A CN201810795696 A CN 201810795696A CN 108985233 A CN108985233 A CN 108985233A
- Authority
- CN
- China
- Prior art keywords
- tracing area
- vehicle
- sub
- image
- district
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- 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
- 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
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/25—Determination of region of interest [ROI] or a volume of interest [VOI]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/46—Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features
- G06V10/462—Salient features, e.g. scale invariant feature transforms [SIFT]
Abstract
The invention discloses one kind to be based on the relevant high-precision wireless vehicle tracking of digital picture, specifically comprise the following steps: to carry out vehicle detection using classical AdaBoost+Haar feature detection algorithm, using the vehicle image region of indicia framing label as initial tracing area；Then, final tracing area is intercepted；Final tracing area after interception is zoomed into fixed size, current frame image is zoomed in and out using same scaling；Using the final tracing area after scaling as reference picture sub-district, centered on the central point of reference picture sub-district, it is area-of-interest that specific dimensions size is chosen in current frame image, carries out the other search of sub-pixel using Digital Image Correlation Method and matches, finds target image sub-district.This invention ensures that real-time, improves the precision of vehicle tracking, more accurate early warning is realized, improve drive safety.
Description
Technical field
The present invention relates to vehicle tracking processing technology fields, specially a kind of to be based on the relevant high-precision vehicle of digital picture
Tracking.
Background technique
With the development of China's communication, car ownership is continuously increased, and in this bring traffic safety thing
Therefore especially rear-end collision is high, according to statistics, in all traffic accidents of China, 34.29% is rear-end impact.
A large amount of rear-end collisions caused by driver attention does not concentrate have been identified as the main safety problem of automobile.Therefore accurate
Early warning to reduce traffic accident play an important role.
The early warning of accurate vehicle tracking and collision is closely related, and on the one hand inaccurate tracking can generate alarm too late
When, it can not effectively avoid traffic accident, on the other hand can generate wrong report, reduce the comfort of driving of driver.Traditional
Meanshift track algorithm is simple, and robustness is good, but works as background since window width size remains unchanged during tracking
Complexity, when target scale is varied, tracking just be will fail, and cannot achieve subsequent calculating；CMT algorithm is a kind of based on feature
The tracking of point has used classical optical flow method, can track any scene any object, and realize tracking box with target ruler
The variation of degree and change, but in practical application, the algorithm takes a long time, the tracking inaccuracy in remote tracking, Wu Fayou
What is chosen when imitating and realize remote vehicle anti-collision warning, while calculating target size scaling is in all characteristic point calculated results
Value is scale value, and there are certain errors.
Summary of the invention
The purpose of the present invention is to provide one kind to be based on the relevant high-precision wireless vehicle tracking of digital picture, on solving
State the problem of proposing in background technique.
To achieve the above object, the invention provides the following technical scheme:
One kind being based on the relevant high-precision wireless vehicle tracking of digital picture, specifically comprises the following steps:
S1, acquisition positive sample and negative sample；All samples are pre-processed；Extraction calculates characteristic value；Using all
The characteristic value of sample is trained, and obtains vehicle detection classifier；
S2, by video frame decoding to be tested, current frame image is pre-processed, extracts and calculates the feature of current frame image
Characteristic value is inputted vehicle detection classifier, the vehicle image region for meeting similarity indices is marked with indicia framing by value,
Using the vehicle image region of indicia framing label as initial tracing area；
S3, the final tracing area of interception, final tracing area is centered on the center of tracing area initial in step S2, most
Whole tracing area side length is 0.8 times of initial tracing area side length；
S4, the final tracing area after interception is zoomed into fixed size, using same scaling to current
Frame image zooms in and out；
S5, using scale after final tracing area as reference picture sub-district, centered on the central point of reference picture sub-district,
It is area-of-interest that specific dimensions size is chosen in current frame image, carries out sub-pix rank using Digital Image Correlation Method
Search matching, find target image sub-district, when ZNSSD be greater than or equal to setting threshold value when, expression track successfully, with label
Frame re-starts label to target image sub-district；When being less than threshold value, indicates tracking failure, continue to execute S2.
As a further solution of the present invention: positive sample selects the vehicle of different automobile types, angle and distance in the step S1
Head-tail picture, negative sample select the non-vehicle picture under driving environment.
As a further solution of the present invention: the concrete operations of the step S1 are as follows: all samples are subjected to image grayscale
Change processing, size are normalized to 24 × 24, and by integrogram, extraction calculates Like-Fenton Oxidation value, uses AdaBoost algorithm
It is trained, obtains AdaBoost vehicle detection classifier.
As a further solution of the present invention: in the step S4: the final tracing area after interception is zoomed to 23 ×
23 sizes.
As a further solution of the present invention: in the step S5: it is sense that 35 × 35 sizes are chosen in current frame image
Interest region.
Compared with prior art, the beneficial effects of the present invention are: being tracking target with the vehicle detected, in target area
Surrounding carries out the other search matching of sub-pixel, and constantly updates the indicia framing size of tracking, is guaranteeing the same of algorithm real-time
When improve the precision of vehicle tracking, realize more accurate early warning, improve drive safety；In addition, the method for the present invention due to
The other search matching of sub-pixel can be achieved, therefore reduce the dependence to high resolution camera, reduce cost, have good
Application value.
Detailed description of the invention
Fig. 1 is the flow diagram of the method for the present invention；
Fig. 2 is the result figure using CMT algorithm keeps track；
Fig. 3 is the result figure of the method for the present invention tracking.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, it is obtained by those of ordinary skill in the art without making creative efforts every other
Embodiment shall fall within the protection scope of the present invention.
Fig. 1~3 are please referred to, the present invention provides a kind of technical solution: one kind is based on the relevant high-precision vehicle of digital picture
Tracking specifically comprises the following steps:
S1, acquisition positive sample and negative sample；All samples are pre-processed；Extraction calculates characteristic value；Using all
The characteristic value of sample is trained, and obtains vehicle detection classifier；
S2, by video frame decoding to be tested, current frame image is pre-processed, extracts and calculates the feature of current frame image
Characteristic value is inputted vehicle detection classifier, the vehicle image region for meeting similarity indices is marked with indicia framing by value,
Using the vehicle image region of indicia framing label as initial tracing area；
S3, the final tracing area of interception, final tracing area is centered on the center of tracing area initial in step S2, most
Whole tracing area side length is 0.8 times of initial tracing area side length；
S4, the final tracing area after interception is zoomed into 23 × 23 sizes, using same scaling to present frame
Image zooms in and out；
S5, using scale after final tracing area as reference picture sub-district, centered on the central point of reference picture sub-district,
It is area-of-interest that 35 × 35 sizes are chosen in current frame image, and it is other to carry out sub-pixel using Digital Image Correlation Method
Search matching, finds target image sub-district, and when ZNSSD is greater than or equal to the threshold value of setting, expression is tracked successfully, uses indicia framing
Label is re-started to target image sub-district；When being less than threshold value, indicates tracking failure, continue to execute S2.
Wherein, in the step S1 positive sample selection different automobile types, angle and distance vehicle head-tail picture, negative sample
The non-vehicle picture under driving environment is selected, such as: tree, road surface, sky without vehicle etc..
The concrete operations of the step S1 are as follows: all samples are subjected to image gray processing processing, size is normalized to 24 ×
24, by integrogram, extraction is calculated Like-Fenton Oxidation value, is trained using AdaBoost algorithm, obtains AdaBoost vehicle
Detection classifier.
Vehicle detection is carried out using classical AdaBoost+Haar feature detection algorithm first, is made with the vehicle detected
For initial tracing area；Secondly, the background for considering that the vehicle region boundary part detected includes is more, therefore to detect
Vehicle region center centered on, side length is long 0.8 times of primary side as final vehicle tracking region, in this way can be effective
Background interference is reduced, while reducing calculation amount.
From Fig. 2 and Fig. 3 as can be seen that the method for the present invention tracking indicia framing scaling be able to carry out smoothly it is high-precision
Degree tracking, and CMT algorithm influenced to fluctuate in scaling by picture noise during tracking it is larger, especially remote
Place can not effectively realize and accurately track, while when object angle change is big can not find characteristic point and tracking is easy to cause to fail.Separately
Outside, the real-time of the method for the present invention tracking is preferable, tracking time-consuming is greatly reduced compared with CMT algorithm, while guaranteeing precision
Reduce time-consuming.
It is obvious to a person skilled in the art that invention is not limited to the details of the above exemplary embodiments, Er Qie
In the case where without departing substantially from spirit or essential attributes of the invention, the present invention can be realized in other specific forms.Therefore, no matter
From the point of view of which point, the present embodiments are to be considered as illustrative and not restrictive, and the scope of the present invention is by appended power
Benefit requires rather than above description limits, it is intended that all by what is fallen within the meaning and scope of the equivalent elements of the claims
Variation is included within the present invention.Any reference signs in the claims should not be construed as limiting the involved claims.
Claims (5)
1. one kind is based on the relevant high-precision wireless vehicle tracking of digital picture, it is characterised in that: specifically comprise the following steps:
S1, acquisition positive sample and negative sample；All samples are pre-processed；Extraction calculates characteristic value；Utilize all samples
Characteristic value be trained, obtain vehicle detection classifier；
S2, by video frame decoding to be tested, current frame image is pre-processed, extracts and calculates the characteristic value of current frame image,
Characteristic value is inputted into vehicle detection classifier, the vehicle image region for meeting similarity indices is marked with indicia framing, it will
The vehicle image region of indicia framing label is as initial tracing area；
S3, the final tracing area of interception, final tracing area centered on the center of tracing area initial in step S2, finally with
0.8 times of a length of initial tracing area side length of track regional edge；
S4, the final tracing area after interception is zoomed into fixed size, using same scaling to present frame figure
As zooming in and out；
S5, using scale after final tracing area as reference picture sub-district, centered on the central point of reference picture sub-district, working as
It is area-of-interest that specific dimensions size is chosen in prior image frame, carries out that sub-pixel is other searches using Digital Image Correlation Method
Rope matching, finds target image sub-district, and when ZNSSD is greater than or equal to the threshold value of setting, expression is tracked successfully, with indicia framing pair
Target image sub-district re-starts label；When being less than threshold value, indicates tracking failure, continue to execute S2.
2. according to claim 1 a kind of based on the relevant high-precision wireless vehicle tracking of digital picture, it is characterised in that:
Positive sample selects the vehicle head-tail picture of different automobile types, angle and distance in the step S1, and negative sample selects driving environment
Under non-vehicle picture.
3. according to claim 1 a kind of based on the relevant high-precision wireless vehicle tracking of digital picture, it is characterised in that:
The concrete operations of the step S1 are as follows: all samples are subjected to image gray processing processing, size is normalized to 24 × 24, by product
Component, extraction are calculated Like-Fenton Oxidation value, are trained using AdaBoost algorithm, and the classification of AdaBoost vehicle detection is obtained
Device.
4. according to claim 1 a kind of based on the relevant high-precision wireless vehicle tracking of digital picture, it is characterised in that:
In the step S4: the final tracing area after interception is zoomed to 23 × 23 sizes.
5. according to claim 1 a kind of based on the relevant high-precision wireless vehicle tracking of digital picture, it is characterised in that:
In the step S5: it is area-of-interest that 35 × 35 sizes are chosen in current frame image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810795696.XA CN108985233B (en) | 2018-07-19 | 2018-07-19 | High-precision vehicle tracking method based on digital image correlation |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810795696.XA CN108985233B (en) | 2018-07-19 | 2018-07-19 | High-precision vehicle tracking method based on digital image correlation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108985233A true CN108985233A (en) | 2018-12-11 |
| CN108985233B CN108985233B (en) | 2021-08-03 |
Family
ID=64549821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810795696.XA Active CN108985233B (en) | 2018-07-19 | 2018-07-19 | High-precision vehicle tracking method based on digital image correlation |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108985233B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079730A (en) * | 2019-11-20 | 2020-04-28 | 北京云聚智慧科技有限公司 | Method for determining area of sample image in interface image and electronic equipment |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003102505A1 (en) * | 2002-05-30 | 2003-12-11 | Rafael - Armament Development Authority Ltd. | Airborne reconnaissance system |
| US20100310128A1 (en) * | 2009-06-03 | 2010-12-09 | Athanasios Iliopoulos | System and Method for Remote Measurement of Displacement and Strain Fields |
| US20130070048A1 (en) * | 2011-09-21 | 2013-03-21 | National Applied Research Laboratories | Formation Apparatus Using Digital Image Correlation |
| CN104809437A (en) * | 2015-04-28 | 2015-07-29 | 无锡赛睿科技有限公司 | Real-time video based vehicle detecting and tracking method |
| CN107038423A (en) * | 2017-04-20 | 2017-08-11 | 常州智行科技有限公司 | A kind of vehicle is detected and tracking in real time |
| WO2017200395A1 (en) * | 2016-05-18 | 2017-11-23 | Auckland Uniservices Limited | Image registration method |
| CN107633526A (en) * | 2017-09-04 | 2018-01-26 | 腾讯科技（深圳）有限公司 | A kind of image trace point acquisition methods and equipment, storage medium |
-
2018
- 2018-07-19 CN CN201810795696.XA patent/CN108985233B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2003102505A1 (en) * | 2002-05-30 | 2003-12-11 | Rafael - Armament Development Authority Ltd. | Airborne reconnaissance system |
| US20100310128A1 (en) * | 2009-06-03 | 2010-12-09 | Athanasios Iliopoulos | System and Method for Remote Measurement of Displacement and Strain Fields |
| US20130070048A1 (en) * | 2011-09-21 | 2013-03-21 | National Applied Research Laboratories | Formation Apparatus Using Digital Image Correlation |
| CN104809437A (en) * | 2015-04-28 | 2015-07-29 | 无锡赛睿科技有限公司 | Real-time video based vehicle detecting and tracking method |
| WO2017200395A1 (en) * | 2016-05-18 | 2017-11-23 | Auckland Uniservices Limited | Image registration method |
| CN107038423A (en) * | 2017-04-20 | 2017-08-11 | 常州智行科技有限公司 | A kind of vehicle is detected and tracking in real time |
| CN107633526A (en) * | 2017-09-04 | 2018-01-26 | 腾讯科技（深圳）有限公司 | A kind of image trace point acquisition methods and equipment, storage medium |
Non-Patent Citations (5)
| Title |
|---|
| MING-DONG YANG等: "Study of image matching algorithm and sub-pixel fitting algorithm in target tracking", 《SELECTED PAPERS FROM CONFERENCES OF THE PHOTOELECTRONIC TECHNOLOGY COMMITTEE OF THE CHINESE SOCIETY OF ASTRONAUTICS 2014, PART I》 * |
| PAN B等: "Fast, Robust and Accurate Digital Image Correlation Calculation Without Redundant Computations", 《EXPERIMENTAL MECHANICS》 * |
| SAMUEL KUCERA等: "using particle image velocimetry for road vehicle tracking and performance monitoring", 《SAE TECHNICAL PAPER》 * |
| 张华俊: "基于图像特征匹配技术的数字图像相关法研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
| 梅华丰等: "利用数字图像相关方法实现高速相机序列影像运动目标的快速跟踪", 《地理信息世界》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111079730A (en) * | 2019-11-20 | 2020-04-28 | 北京云聚智慧科技有限公司 | Method for determining area of sample image in interface image and electronic equipment |
| CN111079730B (en) * | 2019-11-20 | 2023-12-22 | 北京云聚智慧科技有限公司 | Method for determining area of sample graph in interface graph and electronic equipment |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108985233B (en) | 2021-08-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108388641B (en) | Traffic facility map generation method and system based on deep learning | |
| CN106919915B (en) | Map road marking and road quality acquisition device and method based on ADAS system | |
| CN110738121A (en) | front vehicle detection method and detection system | |
| Liu et al. | Rear vehicle detection and tracking for lane change assist | |
| Borkar et al. | A layered approach to robust lane detection at night | |
| Kaur et al. | Lane detection techniques: A review | |
| CN103021177B (en) | Method and system for processing traffic monitoring video image in foggy day | |
| Haque et al. | A computer vision based lane detection approach | |
| KR20180042254A (en) | Systems and methods for object tracking | |
| Ohgushi et al. | Road obstacle detection method based on an autoencoder with semantic segmentation | |
| CN103034843B (en) | Method for detecting vehicle at night based on monocular vision | |
| CN103902976A (en) | Pedestrian detection method based on infrared image | |
| CN104463903A (en) | Pedestrian image real-time detection method based on target behavior analysis | |
| Liu et al. | Development of a vision-based driver assistance system with lane departure warning and forward collision warning functions | |
| CN103488975A (en) | Zebra crossing real-time detection method based in intelligent driving | |
| CN113091757B (en) | Map generation method and device | |
| Raviteja et al. | Advanced driver assitance system (ADAS) | |
| Pepperell et al. | Towards persistent visual navigation using smart | |
| CN108399398A (en) | A kind of pilotless automobile obstacle recognition detection method based on deep learning | |
| CN104881661A (en) | Vehicle detection method based on structure similarity | |
| Qing et al. | A novel particle filter implementation for a multiple-vehicle detection and tracking system using tail light segmentation | |
| CN105512641A (en) | Method for using laser radar scanning method to calibrate dynamic pedestrians and vehicles in video in snowing or raining state | |
| CN104200226A (en) | Particle filtering target tracking method based on machine learning | |
| CN108985233A (en) | One kind being based on the relevant high-precision wireless vehicle tracking of digital picture | |
| CN103927523B (en) | Fog level detection method based on longitudinal gray features |
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 |