CN108985233B - High-precision vehicle tracking method based on digital image correlation - Google Patents
High-precision vehicle tracking method based on digital image correlationDownload PDFInfo
- Publication number
- CN108985233B CN108985233BCN201810795696.XACN201810795696ACN108985233BCN 108985233 BCN108985233 BCN 108985233BCN 201810795696 ACN201810795696 ACN 201810795696ACN 108985233 BCN108985233 BCN 108985233B
- Authority
- CN
- China
- Prior art keywords
- vehicle
- area
- image
- tracking
- tracking area
- 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.)
- Expired - Fee Related
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; 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 OR CALCULATING; 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 OR CALCULATING; 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 OR CALCULATING; 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]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Evolutionary Computation (AREA)
- Evolutionary Biology (AREA)
- General Engineering & Computer Science (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Closed-Circuit Television Systems (AREA)
- Traffic Control Systems (AREA)
Abstract
Description
Technical Field
The invention relates to the technical field of vehicle tracking processing, in particular to a high-precision vehicle tracking method based on digital image correlation.
Background
With the development of the traffic industry of China, the quantity of automobile reserves is continuously increased, and road traffic safety accidents, particularly rear-end collisions, caused by the automobile reserves are high, and according to statistics, 34.29 percent of all highway traffic accidents in China are rear-end collisions. A number of rear-end accidents due to driver inattention have been identified as a major safety issue in automobiles. Therefore, accurate early warning plays an important role in reducing traffic accidents.
Accurate vehicle tracking is closely related to collision early warning, inaccurate tracking can produce the warning untimely on the one hand, can't effectively avoid the traffic accident, and on the other hand can produce the false positive, reduces driver's driving travelling comfort. The traditional meanshift tracking algorithm is simple and good in robustness, but in the tracking process, as the width of a window is kept unchanged, when the background is complex and the target scale is changed, tracking fails, and subsequent calculation cannot be realized; the CMT algorithm is a tracking method based on feature points, a classical optical flow method is used, any object in any scene can be tracked, the tracking frame is changed along with the change of the target dimension, the algorithm consumes a long time in practical application, the tracking is inaccurate in remote tracking, the collision early warning of a remote vehicle cannot be effectively realized, and meanwhile, the median of calculation results of all the feature points is selected as a scaling value when the target dimension is calculated and scaled, and a certain error exists.
Disclosure of Invention
The invention aims to provide a high-precision vehicle tracking method based on digital image correlation, so as to solve the problems in the background technology.
In order to achieve the purpose, the invention provides the following technical scheme:
a high-precision vehicle tracking method based on digital image correlation specifically comprises the following steps:
s1, collecting a positive sample and a negative sample; preprocessing all samples; extracting and calculating a characteristic value; training by using the characteristic values of all samples to obtain a vehicle detection classifier;
s2, unframing the video to be tested, preprocessing the current frame image, extracting and calculating the characteristic value of the current frame image, inputting the characteristic value into a vehicle detection classifier, marking the vehicle image area meeting the similarity index by a marking frame, and taking the vehicle image area marked by the marking frame as an initial tracking area;
s3, intercepting a final tracking area, wherein the final tracking area takes the center of the initial tracking area in the step S2 as the center, and the side length of the final tracking area is 0.8 times of that of the initial tracking area;
s4, zooming the intercepted final tracking area to a fixed size, and zooming the current frame image by adopting the same zooming ratio;
s5, taking the zoomed final tracking area as a reference image subarea and the central point of the reference image subarea as the center, selecting a specific size as an interesting area in the current image, performing subpixel-level search and matching by adopting a digital image correlation method, searching for a target image subarea, indicating that the tracking is successful when ZNSDS is larger than or equal to a set threshold value, and re-marking the target image subarea by using a marking frame; when the value is less than the threshold value, it indicates that the tracking has failed, and the process proceeds to S2.
As a further scheme of the invention: in the step S1, the positive sample selects pictures of the head and the tail of the vehicle with different vehicle types, angles and distances, and the negative sample selects a non-vehicle picture in the driving environment.
As a further scheme of the invention: the specific operation of step S1 is: and performing image graying processing on all samples, normalizing the size to be 24 multiplied by 24, extracting and calculating a Haar-like characteristic value by means of an integrogram, and training by using an AdaBoost algorithm to obtain the AdaBoost vehicle detection classifier.
As a further scheme of the invention: in the step S4: the truncated final tracking area is scaled to 23 x 23 size.
As a further scheme of the invention: in the step S5: and selecting a 35 multiplied by 35 size as a region of interest in the current frame image.
Compared with the prior art, the invention has the beneficial effects that: the method has the advantages that the detected vehicle is taken as a tracking target, sub-pixel level searching and matching are carried out around the target area, and the size of a tracked marking frame is continuously updated, so that the accuracy of vehicle tracking is improved while the real-time performance of an algorithm is ensured, more accurate early warning is realized, and the driving safety is improved; in addition, the method can realize the search and matching at the sub-pixel level, thereby reducing the dependence on a high-resolution camera, reducing the cost and having good application value.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a diagram of the results tracked using the CMT algorithm;
FIG. 3 is a graph of the results of the tracking by the method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 3, the present invention provides a technical solution: a high-precision vehicle tracking method based on digital image correlation specifically comprises the following steps:
s1, collecting a positive sample and a negative sample; preprocessing all samples; extracting and calculating a characteristic value; training by using the characteristic values of all samples to obtain a vehicle detection classifier;
s2, unframing the video to be tested, preprocessing the current frame image, extracting and calculating the characteristic value of the current frame image, inputting the characteristic value into a vehicle detection classifier, marking the vehicle image area meeting the similarity index by a marking frame, and taking the vehicle image area marked by the marking frame as an initial tracking area;
s3, intercepting a final tracking area, wherein the final tracking area takes the center of the initial tracking area in the step S2 as the center, and the side length of the final tracking area is 0.8 times of that of the initial tracking area;
s4, scaling the intercepted final tracking area to 23 multiplied by 23 size, and scaling the current frame image by the same scaling;
s5, taking the zoomed final tracking area as a reference image subarea, taking the central point of the reference image subarea as the center, selecting 35 x 35 size as an interesting area in the current image, adopting a digital image correlation method to search and match at a sub-pixel level, searching a target image subarea, when ZNSD is larger than or equal to a set threshold value, indicating that the tracking is successful, and marking the target image subarea again by using a marking frame; when the value is less than the threshold value, it indicates that the tracking has failed, and the process proceeds to S2.
In step S1, the positive sample selects pictures of the head and the tail of the vehicle with different models, angles and distances, and the negative sample selects non-vehicle pictures in the driving environment, such as: not including trees, roads, sky, etc. of the vehicle.
The specific operation of step S1 is: and performing image graying processing on all samples, normalizing the size to be 24 multiplied by 24, extracting and calculating a Haar-like characteristic value by means of an integrogram, and training by using an AdaBoost algorithm to obtain the AdaBoost vehicle detection classifier.
Firstly, vehicle detection is carried out by using a classical AdaBoost + Haar characteristic detection algorithm, and a detected vehicle is used as an initial tracking area; secondly, considering that the detected vehicle area boundary part contains more backgrounds, the center of the detected vehicle area is taken as the center, and the side length is 0.8 times of the original side length to be taken as the final vehicle tracking area, so that the background interference can be effectively reduced, and the calculation amount is reduced.
As can be seen from FIG. 2 and FIG. 3, the scale scaling of the mark frame tracked by the method of the present invention can perform stable and high-precision tracking, and the CMT algorithm is affected by image noise in the tracking process and has large fluctuation on the scale scaling, especially the CMT algorithm cannot effectively realize accurate tracking at a long distance, and meanwhile, when the angle change of an object is large, a tracking failure is easily caused because a feature point cannot be found. In addition, the method has better tracking real-time performance, greatly reduces the tracking time consumption compared with a CMT algorithm, ensures the precision and reduces the time consumption.
It will be evident to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, and that the present invention may be embodied in other specific forms without departing from the spirit or essential attributes thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Claims (5)
1. A high-precision vehicle tracking method based on digital image correlation is characterized in that: the method specifically comprises the following steps:
s1, collecting a positive sample and a negative sample; preprocessing all samples; extracting and calculating a characteristic value; training by using the characteristic values of all samples to obtain a vehicle detection classifier;
s2, unframing the video to be tested, preprocessing the current frame image, extracting and calculating the characteristic value of the current frame image, inputting the characteristic value into a vehicle detection classifier, marking the vehicle image area meeting the similarity index by a marking frame, and taking the vehicle image area marked by the marking frame as an initial tracking area;
s3, intercepting a final tracking area, wherein the final tracking area takes the center of the initial tracking area in the step S2 as the center, and the side length of the final tracking area is 0.8 times of that of the initial tracking area;
s4, zooming the intercepted final tracking area to a fixed size, and zooming the current frame image by adopting the same zooming ratio;
s5, taking the zoomed final tracking area as a reference image subarea and the central point of the reference image subarea as the center, selecting a specific size as an interesting area in the current image, performing subpixel-level search and matching by adopting a digital image correlation method, searching for a target image subarea, indicating that the tracking is successful when ZNSDS is larger than or equal to a set threshold value, and re-marking the target image subarea by using a marking frame; when the value is less than the threshold value, it indicates that the tracking has failed, and the process proceeds to S2.
2. The digital image correlation-based high-precision vehicle tracking method according to claim 1, characterized in that: in the step S1, the positive sample selects pictures of the head and the tail of the vehicle with different vehicle types, angles and distances, and the negative sample selects a non-vehicle picture in the driving environment.
3. The digital image correlation-based high-precision vehicle tracking method according to claim 1, characterized in that: the specific operation of step S1 is: and performing image graying processing on all samples, normalizing the size to be 24 multiplied by 24, extracting and calculating a Haar-like characteristic value by means of an integrogram, and training by using an AdaBoost algorithm to obtain the AdaBoost vehicle detection classifier.
4. The digital image correlation-based high-precision vehicle tracking method according to claim 1, characterized in that: in the step S4: the truncated final tracking area is scaled to 23 x 23 size.
5. The digital image correlation-based high-precision vehicle tracking method according to claim 1, characterized in that: in the step S5: and selecting a 35 multiplied by 35 size as a region of interest in the current frame image.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810795696.XACN108985233B (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.XACN108985233B (en) | 2018-07-19 | 2018-07-19 | High-precision vehicle tracking method based on digital image correlation |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108985233A CN108985233A (en) | 2018-12-11 |
| CN108985233Btrue CN108985233B (en) | 2021-08-03 |
Family
ID=64549821
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810795696.XAExpired - Fee RelatedCN108985233B (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) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
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
- 2018-07-19CNCN201810795696.XApatent/CN108985233B/ennot_activeExpired - Fee Related
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 |
|---|
| Fast, Robust and Accurate Digital Image Correlation Calculation Without Redundant Computations;Pan B等;《Experimental Mechanics》;20130207;第53卷(第7期);1277-1289页* |
| Study of image matching algorithm and sub-pixel fitting algorithm in target tracking;Ming-dong Yang等;《Selected Papers from Conferences of the Photoelectronic Technology Committee of the Chinese Society of Astronautics 2014, Part I》;20150304;第9521卷;95211M页* |
| using particle image velocimetry for road vehicle tracking and performance monitoring;Samuel Kucera等;《SAE Technical Paper》;20110412;第1卷;0279页* |
| 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;梅华丰等;《地理信息世界》;20131031;第20卷(第5期);51-58页* |
| 基于图像特征匹配技术的数字图像相关法研究;张华俊;《中国优秀硕士学位论文全文数据库 信息科技辑》;20140815(第8期);I138-1473页* |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108985233A (en) | 2018-12-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11643076B2 (en) | Forward collision control method and apparatus, electronic device, program, and medium | |
| Mahaur et al. | Road object detection: a comparative study of deep learning-based algorithms | |
| CN101996410B (en) | Method and system of detecting moving object under dynamic background | |
| CN110069986B (en) | A hybrid model-based traffic signal recognition method and system | |
| CN114359181B (en) | Intelligent traffic target fusion detection method and system based on image and point cloud | |
| Pavlic et al. | Classification of images in fog and fog-free scenes for use in vehicles | |
| Haque et al. | A computer vision based lane detection approach | |
| US9292750B2 (en) | Method and apparatus for detecting traffic monitoring video | |
| WO2020151172A1 (en) | Moving object detection method and apparatus, computer device, and storage medium | |
| CN109034047A (en) | A kind of method for detecting lane lines and device | |
| CN103413444A (en) | Traffic flow surveying and handling method based on unmanned aerial vehicle high-definition video | |
| CN111814593B (en) | Traffic scene analysis method and equipment and storage medium | |
| Gluhaković et al. | Vehicle detection in the autonomous vehicle environment for potential collision warning | |
| CN107038423B (en) | Vehicle real-time detection and tracking method | |
| CN112927283A (en) | Distance measuring method and device, storage medium and electronic equipment | |
| Sang et al. | A robust lane detection algorithm adaptable to challenging weather conditions | |
| Li et al. | Improved YOLOv5s algorithm for small target detection in UAV aerial photography | |
| CN110154896B (en) | Method and equipment for detecting obstacle | |
| CN111422203B (en) | Driving behavior evaluation method and device | |
| Wang et al. | An improved hough transform method for detecting forward vehicle and lane in road | |
| Vajak et al. | A rethinking of real-time computer vision-based lane detection | |
| Qing et al. | A novel particle filter implementation for a multiple-vehicle detection and tracking system using tail light segmentation | |
| CN108985233B (en) | High-precision vehicle tracking method based on digital image correlation | |
| Zhang et al. | Real-time lane detection method based on region of interest | |
| CN118429941A (en) | Traffic signal lamp identification 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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee | Granted publication date:20210803 |