CN112419364A - Target tracking method and system based on image feature matching - Google Patents
Target tracking method and system based on image feature matching Download PDFInfo
- Publication number
- CN112419364A CN112419364A CN202011232734.4A CN202011232734A CN112419364A CN 112419364 A CN112419364 A CN 112419364A CN 202011232734 A CN202011232734 A CN 202011232734A CN 112419364 A CN112419364 A CN 112419364A
- Authority
- CN
- China
- Prior art keywords
- target
- frame
- value
- pixel value
- targets
- 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
- 238000000034 method Methods 0.000 title claims abstract description 28
- 238000001514 detection method Methods 0.000 claims abstract description 22
- 238000006073 displacement reaction Methods 0.000 claims description 21
- 230000008030 elimination Effects 0.000 claims description 16
- 238000003379 elimination reaction Methods 0.000 claims description 16
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000005259 measurement Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 6
- 238000003384 imaging method Methods 0.000 description 3
- 238000005516 engineering process Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004907 flux Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000002123 temporal effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/20—Analysis of motion
- G06T7/246—Analysis of motion using feature-based methods, e.g. the tracking of corners or segments
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/136—Segmentation; Edge detection involving thresholding
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/194—Segmentation; Edge detection involving foreground-background segmentation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/60—Analysis of geometric attributes
- G06T7/66—Analysis of geometric attributes of image moments or centre of gravity
-
- 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
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10016—Video; Image sequence
-
- 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
-
- 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
- G06V2201/08—Detecting or categorising vehicles
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Geometry (AREA)
- Image Analysis (AREA)
Abstract
The invention discloses a target tracking method and a target tracking system based on image feature matching, which relate to the technical field of image processing and solve the technical problem that a plurality of targets cannot be accurately captured in a high-resolution image in real time. The method and the system can accurately capture a plurality of targets in a high-resolution image in real time, and realize real-time target detection.
Description
Technical Field
The present disclosure relates to the field of image processing technologies, and in particular, to a target tracking method and system based on image feature matching.
Background
The real-time speed information collection plays a very important role in the Traffic flow information collection, and the realization of an Intelligent Traffic System (ITS) can not separate the real-time speed information collection. The speed measurement by hardware mainly includes laser speed measurement, radar speed measurement, and the like, but it requires professional equipment, is expensive to install and maintain, and has limitations in application. The speed measurement through software mainly comprises video-based speed measurement, and the target detection is carried out on the traffic flow video by adopting an image processing technology, so that the aim of collecting various traffic information is fulfilled, and rich traffic flow information is provided.
Currently, methods for object detection mainly include a frame difference method and an optical flow method. The frame difference method is to extract a motion region of a moving object in an image by closed-valued based on temporal difference of pixels between adjacent frames of an image sequence and can eliminate interference of an unstable factor, but the frame difference method is limited in that it cannot detect a stationary and slowly moving object. The optical flow method reflects the change of pixel points in the time domain of the imaging plane coordinate system by calculating the optical flux value of the pixel points of the video image, then calculates the corresponding relation between adjacent frames, and calculates the tracked image displacement target and other information. However, the basic assumption of the optical flow method is that the luminance is constant and the moving time is short, so the optical flow method is not applicable when the light changes or the time interval is too long, most of moving object tracking is based on the principles of motion estimation, probability distribution, feature matching and the like, the calculation amount of the optical flow method is large, and the real-time requirement of video speed detection cannot be met. Therefore, the prior art cannot accurately capture a plurality of targets in a high-resolution image in real time, and cannot meet the real-time requirement of target detection.
Disclosure of Invention
The present disclosure provides a target tracking method and system based on image feature matching, which aims to accurately capture a plurality of targets in a high-resolution image in real time and realize real-time target detection.
The technical purpose of the present disclosure is achieved by the following technical solutions:
a target tracking method based on image feature matching comprises the following steps:
acquiring continuous video images, and cutting the video images into continuous first frame pictures;
background elimination is carried out on the first frame of picture to extract a target image, and a second frame of picture is obtained;
comparing the pixel value of the second frame of picture with the historical value of the pixel value, and if the difference value between the pixel value and the historical value of the pixel value does not exceed a first preset threshold value, storing the pixel value into a potential category; if the difference between the pixel value and the historical value exceeds a first preset threshold, storing the pixel value into a potential background;
judging whether the tracked targets are the same target or not according to the potential categories and the centroid distance between the targets;
and if the target is the same target, calculating the actual displacement of the target according to the pixel displacement of the pixel values in the potential category.
Further, the determining whether the tracked targets are the same target according to the potential categories and the centroid distance between the targets includes:
determining a target detection frame in an adjacent frame of the second frame picture;
calculating the centroid distance of the target detection frames of the adjacent frames, judging whether the centroid distance is within a second preset threshold value, if the centroid distance is within the second threshold value, the tracked targets are the same targets, and if the centroid distance is not within the second threshold value, judging that the tracked targets are invalid.
Further, the background elimination method is based on KNN algorithm.
A target tracking system based on image feature matching, comprising:
the acquisition module acquires continuous video images and cuts the video images into continuous first frame pictures;
the background elimination module is used for eliminating the background of the first frame of picture to extract a target image so as to obtain a second frame of picture;
the comparison module compares the pixel value of the second frame of picture with the historical value of the pixel value, and stores the pixel value into a potential category if the difference value between the pixel value and the historical value is within a first preset threshold value; if the difference between the pixel value and the historical value is not within a first preset threshold, storing the pixel value into a potential background;
the judging module judges whether the tracked targets are the same target or not according to the potential categories and the centroid distance between the targets;
and the calculation module is used for calculating the actual displacement of the target according to the pixel displacement of the pixel values in the potential category if the target is the same target.
Further, the determining module is further configured to: determining a target detection frame in an adjacent frame of the second frame picture;
calculating the centroid distance of the target detection frames of the adjacent frames, judging whether the centroid distance is within a second preset threshold value, if the centroid distance is within the second threshold value, the tracked targets are the same targets, and if the centroid distance is not within the second threshold value, judging that the tracked targets are invalid.
The beneficial effect of this disclosure lies in: the target tracking method and system based on image feature matching, which are disclosed by the disclosure, perform background elimination on a collected video frame to obtain a target image frame, compare a pixel value of the target image frame with a historical value of the target image frame, and store the pixel value into a potential category if a difference value between the pixel value and the historical value is within a first preset threshold value. In addition, whether the tracked targets are the same target or not is judged according to the potential categories and the centroid distance between the targets; and if the target is the same target, calculating the actual displacement of the target according to the pixel displacement of the pixel values in the potential category. The accurate and stable background image can be extracted through background elimination, the target image frame is accurately separated from the original frame through the background image, whether the tracked target is the same target or not is judged through the centroid distance, and the actual displacement of the target is finally calculated. The method and the system can accurately capture a plurality of targets in a high-resolution image in real time, and realize real-time target detection.
Drawings
FIG. 1 is a flow chart of the disclosed method;
FIG. 2 is a schematic view of the disclosed system;
fig. 3 is a schematic view of a camera imaging model.
Detailed Description
The technical scheme of the disclosure will be described in detail with reference to the accompanying drawings. In the description of the present disclosure, it is to be understood that the terms "first" and "second" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or implying any number of technical features indicated, but merely as distinguishing between different components.
Fig. 1 is a flow chart of the method of the present disclosure, as shown in fig. 1, 100: acquiring continuous video images, and cutting the video images into continuous first frame pictures. As a specific embodiment, the aim of preprocessing a video image sequence is fulfilled by acquiring continuous video images, capturing complete and stable vehicle information and then dividing the video images into a plurality of frame images.
101: and carrying out background elimination on the first frame of picture to extract a target image so as to obtain a second frame of picture. As a specific embodiment, the background elimination method is background elimination performed based on a KNN (K-nearest neighbor classification) algorithm, and background modeling is performed based on the KNN algorithm to perform background elimination, so as to realize detection of an object and a contour, thereby obtaining good initial vehicle characteristics.
102: comparing the pixel value of the second frame of picture with the historical value of the pixel value, and if the difference value between the pixel value and the historical value of the pixel value does not exceed a first preset threshold value, storing the pixel value into a potential category; if the difference between a pixel value and its historical value exceeds a first preset threshold, the pixel value is stored into a potential background.
As a specific embodiment, in a scene, for one pixel value of the second frame picture, if a difference between the pixel value and the history value does not exceed a first preset threshold compared with K history values of the pixel, the pixel value and the history value which are indicated to be a match are stored into a potential category, and after comparing all history fingers, if the difference between the pixel value and the history value exceeds the first preset threshold, the pixel point is stored as a potential background point.
103: and judging whether the tracked targets are the same target or not according to the potential categories and the centroid distance between the targets. As a specific embodiment, the feature that the target contour of the adjacent frame does not change much can be obtained from the pixel values of the potential category. In the invention, the centroid represents a vehicle, and whether the selected target is the same vehicle is judged according to the characteristic that the centroid has smaller distance deviation in the adjacent image sequence.
The specific process of judging whether the tracked targets are the same target according to the centroid distance comprises the following steps:
(1) determining a target detection frame in an adjacent frame of the second frame picture;
(2) calculating the centroid distance of the target detection frames of the adjacent frames, judging whether the centroid distance is within a second preset threshold value, if the centroid distance is within the second threshold value, the tracked target is the same target, namely the target detection frames in the adjacent frames track the same vehicle, displaying two points to represent different positions of the same vehicle, and if the centroid distance is not within the second threshold value, judging that the tracked target is invalid.
Generally, if the centroid distance is not within the second threshold, it means that the two target detection frames are not tracking the same vehicle, or the vehicle is entering and exiting the speed measurement area and is unable to display an image, thereby abandoning the distance, i.e., the tracked target is invalid. When the video frame cannot be read, it indicates that the algorithm is completed and the tracking is completed.
104: and if the target is the same target, calculating the actual displacement of the target according to the pixel displacement of the pixel values in the potential category. And acquiring the unique distance of the center of mass through the matching of the distance of the center of mass so as to facilitate the precision comparison.
For the calculation of the vehicle speed depending on the actual displacement of the vehicle, the invention obtains the pixel displacement of the target vehicle, and the actual displacement needs to be obtained through camera calibration, namely, a geometric model of camera imaging is established, so that the mapping relation between the image coordinate and the actual three-dimensional road coordinate system is solved, as shown in fig. 3, and the pixel displacement is converted into the actual displacement. In addition, the vehicle motion time can be directly obtained through the image frame rate, and finally the vehicle running speed can be obtained according to the vehicle motion time and the actual displacement.
Fig. 2 is a schematic diagram of the system of the present disclosure, which includes an acquisition module, a background elimination module, a comparison module, a judgment module, and a calculation module, where the acquisition module is configured to acquire continuous video images and divide the video images into continuous first frame pictures. The background elimination module is used for eliminating the background of the first frame picture to extract a target image, and a second frame picture is obtained. The comparison module is used for comparing the pixel value of the second frame of picture with the historical value of the pixel value, and if the difference value between the pixel value and the historical value does not exceed a first preset threshold value, the pixel value is stored in a potential category; if the difference between a pixel value and its historical value exceeds a first preset threshold, the pixel value is stored into a potential background. The calculation module is used for calculating the actual displacement of the target.
The judging module is used for judging whether the tracked targets are the same target or not according to the potential categories and the centroid distance between the targets; and is also used for: determining a target detection frame in an adjacent frame of the second frame picture; calculating the centroid distance of the target detection frames of the adjacent frames, judging whether the centroid distance is within a second preset threshold value, if the centroid distance is within the second threshold value, the tracked targets are the same targets, and if the centroid distance is not within the second threshold value, judging that the tracked targets are invalid.
The foregoing is an exemplary embodiment of the present disclosure, and the scope of the present disclosure is defined by the claims and their equivalents.
Claims (6)
1. A target tracking method based on image feature matching is characterized by comprising the following steps:
acquiring continuous video images, and cutting the video images into continuous first frame pictures;
background elimination is carried out on the first frame of picture to extract a target image, and a second frame of picture is obtained;
comparing the pixel value of the second frame of picture with the historical value of the pixel value, and if the difference value between the pixel value and the historical value of the pixel value does not exceed a first preset threshold value, storing the pixel value into a potential category; if the difference between the pixel value and the historical value exceeds a first preset threshold, storing the pixel value into a potential background;
judging whether the tracked targets are the same target or not according to the potential categories and the centroid distance between the targets;
and if the target is the same target, calculating the actual displacement of the target according to the pixel displacement of the pixel values in the potential category.
2. The target tracking method based on image feature matching as claimed in claim 1, wherein the determining whether the tracked targets are the same target according to the potential categories and the centroid distance between the targets comprises:
determining a target detection frame in an adjacent frame of the second frame picture;
calculating the centroid distance of the target detection frames of the adjacent frames, judging whether the centroid distance is within a second preset threshold value, if the centroid distance is within the second threshold value, the tracked targets are the same targets, and if the centroid distance is not within the second threshold value, judging that the tracked targets are invalid.
3. The target tracking method based on image feature matching as claimed in claim 1, wherein the background elimination method is based on KNN algorithm.
4. An image feature matching-based target tracking system, comprising:
the acquisition module acquires continuous video images and cuts the video images into continuous first frame pictures;
the background elimination module is used for eliminating the background of the first frame of picture to extract a target image so as to obtain a second frame of picture;
the comparison module compares the pixel value of the second frame of picture with the historical value of the pixel value, and stores the pixel value into a potential category if the difference value between the pixel value and the historical value is within a first preset threshold value; if the difference between the pixel value and the historical value is not within a first preset threshold, storing the pixel value into a potential background;
the judging module judges whether the tracked targets are the same target or not according to the potential categories and the centroid distance between the targets;
and the calculation module is used for calculating the actual displacement of the target according to the pixel displacement of the pixel values in the potential category if the target is the same target.
5. The image feature matching-based target tracking system of claim 1, wherein the determination module is further configured to: determining a target detection frame in an adjacent frame of the second frame picture;
calculating the centroid distance of the target detection frames of the adjacent frames, judging whether the centroid distance is within a second preset threshold value, if the centroid distance is within the second threshold value, the tracked targets are the same targets, and if the centroid distance is not within the second threshold value, judging that the tracked targets are invalid.
6. The image feature matching-based target tracking system of claim 1, wherein the background elimination method is KNN algorithm-based background elimination.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011232734.4A CN112419364A (en) | 2020-11-06 | 2020-11-06 | Target tracking method and system based on image feature matching |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202011232734.4A CN112419364A (en) | 2020-11-06 | 2020-11-06 | Target tracking method and system based on image feature matching |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112419364A true CN112419364A (en) | 2021-02-26 |
Family
ID=74782090
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202011232734.4A Pending CN112419364A (en) | 2020-11-06 | 2020-11-06 | Target tracking method and system based on image feature matching |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112419364A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113673364A (en) * | 2021-07-28 | 2021-11-19 | 上海影谱科技有限公司 | Video violence detection method and device based on deep neural network |
| TWI790078B (en) * | 2021-12-20 | 2023-01-11 | 財團法人工業技術研究院 | Object detection method and object detection system for video |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109118516A (en) * | 2018-07-13 | 2019-01-01 | 高新兴科技集团股份有限公司 | A kind of target is from moving to static tracking and device |
-
2020
- 2020-11-06 CN CN202011232734.4A patent/CN112419364A/en active Pending
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109118516A (en) * | 2018-07-13 | 2019-01-01 | 高新兴科技集团股份有限公司 | A kind of target is from moving to static tracking and device |
Non-Patent Citations (1)
| Title |
|---|
| GENYUAN CHENG等: "Real-time detection of vehicle speed based on video image", 《2020 12TH INTERNATIONAL CONFERENCE ON MEASURING TECHNOLOGY AND MECHATRONICS AUTOMATION (ICMTMA)》, pages 2 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113673364A (en) * | 2021-07-28 | 2021-11-19 | 上海影谱科技有限公司 | Video violence detection method and device based on deep neural network |
| TWI790078B (en) * | 2021-12-20 | 2023-01-11 | 財團法人工業技術研究院 | Object detection method and object detection system for video |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110441791B (en) | Ground obstacle detection method based on forward-leaning 2D laser radar | |
| KR101647370B1 (en) | road traffic information management system for g using camera and radar | |
| EP2426642B1 (en) | Method, device and system for motion detection | |
| CN101344965A (en) | Tracking system based on binocular camera shooting | |
| CN111723778B (en) | Vehicle distance measuring system and method based on MobileNet-SSD | |
| CN105975923B (en) | Method and system for tracking human objects | |
| CN113848545B (en) | Fusion target detection and tracking method based on vision and millimeter wave radar | |
| CN112419364A (en) | Target tracking method and system based on image feature matching | |
| CN110717445A (en) | Front vehicle distance tracking system and method for automatic driving | |
| Cheng et al. | Real-time detection of vehicle speed based on video image | |
| CN112613568B (en) | Target identification method and device based on visible light and infrared multispectral image sequence | |
| JP2011513876A (en) | Method and system for characterizing the motion of an object | |
| CN115166717A (en) | Lightweight target tracking method integrating millimeter wave radar and monocular camera | |
| JP3629935B2 (en) | Speed measurement method for moving body and speed measurement device using the method | |
| CN116862832A (en) | Three-dimensional live-action model-based operator positioning method | |
| KR101161557B1 (en) | The apparatus and method of moving object tracking with shadow removal moudule in camera position and time | |
| JP2002367077A (en) | Device and method for deciding traffic congestion | |
| CN104537690B (en) | One kind is based on the united moving spot targets detection method of maximum time index | |
| CN115512263A (en) | Dynamic visual monitoring method and device for falling object | |
| CN113553958B (en) | Expressway green belt detection method and device | |
| JPH09322153A (en) | Automatic monitor | |
| CN112561930B (en) | System and method for real-time framing of target in video stream | |
| JPH0991439A (en) | Object monitor | |
| CN113516685A (en) | Target tracking method, device, equipment and storage medium | |
| CN115994934B (en) | Data time alignment method and device and domain controller |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20220111 Address after: 2209-c1, No. 19, Erquan East Road, Huizhi enterprise center, Xishan District, Wuxi City, Jiangsu Province, 214000 Applicant after: Wuxi yuspace Intelligent Technology Co.,Ltd. Address before: Room 1101, block C, Kangyuan smart port, No. 50, Jiangdong Street, Jialing, Jianye District, Nanjing City, Jiangsu Province, 210000 Applicant before: Jiangsu Yu Space Technology Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20220920 Address after: Room 1101, block C, Kangyuan smart port, No. 50, Jiangdong Street, Jialing, Jianye District, Nanjing City, Jiangsu Province, 210000 Applicant after: Jiangsu Yu Space Technology Co.,Ltd. Address before: 2209-c1, No. 19, Erquan East Road, Huizhi enterprise center, Xishan District, Wuxi City, Jiangsu Province, 214000 Applicant before: Wuxi yuspace Intelligent Technology Co.,Ltd. |
|
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230830 Address after: 2209-c1, No. 19, Erquan East Road, Huizhi enterprise center, Xishan District, Wuxi City, Jiangsu Province, 214000 Applicant after: Wuxi yuspace Intelligent Technology Co.,Ltd. Address before: Room 1101, block C, Kangyuan smart port, No. 50, Jiangdong Street, Jialing, Jianye District, Nanjing City, Jiangsu Province, 210000 Applicant before: Jiangsu Yu Space Technology Co.,Ltd. |