Multi-athlete tracking, identifying and timing method and system based on smart phone
By integrating global detectors, trackers and SORT algorithms on smartphones, the occlusion, error and accuracy problems of timing systems in multiple athletes in the prior art are solved, and efficient and accurate athlete tracking and timing functions are achieved.
Patent Information
- Application Number
- CN202510134790.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-06
- Publication Date
- 2025-05-30
AI Technical Summary
The existing running competition timing system has problems such as occlusion, timing error and detection accuracy in multiple athlete scenarios.
A multi-athlete tracking and identification timing system based on smartphones is adopted, combining a global detector, input image preprocessing module, post-processing and filtering module, tracker and SORT algorithm to realize real-time identification and tracking of athletes, non-athletes and finish line.
The detection efficiency and accuracy of multi-athletes' timing tracking are improved, ensuring that each target has only one most accurate detection result, avoiding confusion or loss of targets, and achieving accurate recording of athletes' cross-line moments.
Smart Images
Figure CN120070971A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of running competitions, and particularly relates to a multi-athlete tracking, recognition and timing method and system based on a smart phone. Background Art
[0002] In running competitions, automatic timing devices generally use high-speed photography timing systems, RFID chip timing systems or AI video timing systems to record the running results of athletes participating in running competitions and the crossing times of athletes, and judge the ranking of athletes according to the recorded result data.
[0003] Existing timing systems have the following drawbacks: the high-speed photography timing system is prone to being blocked between multiple athletes outside the venue and cannot effectively time; the chip timing system has timing errors when the distances between multiple athletes are close; the timing camera of the AI timing system is easily blocked, affecting timing. All of the above timing systems have the problem that the detection accuracy cannot be guaranteed. Therefore, a multi-athlete tracking, recognition and timing method and system based on a smart phone are needed to solve the above problems. Summary of the Invention
[0004] The purpose of the present invention is to provide a multi-athlete tracking, recognition and timing method and system based on a smart phone to solve the problems raised in the above background art.
[0005] To achieve the above purpose, the present invention provides the following technical solution: a multi-athlete tracking, recognition and timing system based on a smart phone, including a global detector, a tracker and the SORT algorithm. The global detector is connected with an input image preprocessing module, the global detector is connected with a target detection model module, the global detector is connected with a category recognition and positioning module, the global detector is connected with a post-processing and filtering module, the global detector is connected with an athlete identification module, the global detector is connected with a non-athlete identification module, the global detector is connected with a finish line identification module, the finish line identification module is connected with a finish line feature module, the finish line identification module is connected with a position detection module, the tracker and the SORT algorithm are connected with an athlete tracking module, the tracker and the SORT algorithm are connected with a dynamic tracking module for athletes and non-athletes, and the tracker and the SORT algorithm are connected with a finish line detection and position tracking module.
[0006] A further technical solution is that the main function of the global detector is to identify and classify different targets in the picture, such as athletes, non-athletes and the finish line.
[0007] A further technical solution is that the input image preprocessing module can preprocess the image.
[0008] Further technical solution: The target detection model module can process images using a deep learning-based target detection algorithm and can also detect potential targets therefrom.
[0009] Further technical solution: The category recognition and localization module can classify each athlete, non-athlete, and finish line and also provide their bounding boxes to indicate the positions of the targets in the image.
[0010] Further technical solution: The post-processing and filtering module can handle non-maximum suppression and can also remove overlapping bounding boxes to ensure that each target has only one most accurate detection result.
[0011] A multi-athlete tracking, recognition, and timing method based on a smartphone, applied to the multi-athlete tracking, recognition, and timing system based on a smartphone described in any one of the above, includes the following steps:
[0012] S1. The application software is started;
[0013] S2. The mobile phone camera is called;
[0014] S3. A frame of picture of the video collected by the mobile phone is obtained;
[0015] S4. The first-level detector;
[0016] S5. The tracker;
[0017] S6. The bib detector;
[0018] S7. The bib character detector;
[0019] S8. The finish line action determination detector;
[0020] S9. Data integration to form data labels;
[0021] S10. Record the moment when the athlete crosses the finish line;
[0022] S11. Write the data and pictures into the database.
[0023] Compared with the prior art, the beneficial effects of the present invention are:
[0024] In this invention, by setting up a global detector, an input image preprocessing module, a post-processing and filtering module tracker, and the SORT algorithm, the global detector can receive the preprocessed image or video frame after passing through the input image preprocessing module. Through operations such as scaling, normalization, and color space conversion, it can improve the detection efficiency and accuracy of multi-athlete timing tracking. The global detector can cooperate with the post-processing and filtering module to handle non-maximum suppression and remove overlapping bounding boxes to ensure that each target has only one most accurate detection result, thereby guaranteeing the precision of the multi-athlete timing tracking detection result. The tracker and the SORT algorithm can track each athlete in consecutive video frames through the SORT algorithm system. When tracking multiple targets, the tracker and the SORT algorithm can effectively handle the dynamic changes of multiple targets in the scene, including the interaction between athletes and non-athletes, thus avoiding the situation of confusing or losing targets. And through cooperation with the global detector, the tracker and the SORT algorithm can be used to track the change of the finish line in real time. In scenarios such as track races, the finish line usually appears at a specific frame position. The tracker and the SORT algorithm can track the exact position of the finish line and help athletes with position correction, thereby ensuring that each athlete is uniquely identified and tracked;
[0025] In this invention, by setting up a target detection model module and a category recognition and localization module, the target detection model module can process the image using a deep learning-based target detection algorithm and detect potential targets from it. At the same time, in cooperation with the global detector, it can automatically learn the features of various targets in the image through a convolutional neural network to efficiently identify and locate athletes, non-athletes, and the finish line in the picture. The category recognition and localization module can classify each athlete, non-athlete, and the finish line, and also provide their bounding boxes to indicate the position of the target in the image, thereby ensuring the detection accuracy of each athlete, non-athlete, and the finish line. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 It is a schematic structural diagram of the execution process module of the present invention;
[0027] Figure 2 It is a schematic structural diagram of the global detector connection module of the present invention;
[0028] Figure 3 It is a schematic structural diagram of the global detector identification module of the present invention;
[0029] Figure 4 It is a schematic structural diagram of the tracker and SORT algorithm module of the present invention.
[0030] In the figure: 1. Global detector; 2. Input image preprocessing module; 3. Target detection model module; 4. Class recognition and localization module; 5. Post-processing and filtering module; 6. Athlete identification module; 7. Non-athlete identification module; 8. Finish line identification module; 9. Finish line feature module; 10. Position detection module; 11. Tracker and SORT algorithm; 12. Athlete tracking module; 13. Dynamic tracking module for athletes and non-athletes; 14. Finish line detection and position tracking module. Specific implementation mode
[0031] The present invention will be further described below in conjunction with embodiments.
[0032] The following embodiments are used to illustrate the present invention, but cannot be used to limit the protection scope of the present invention. The conditions in the embodiments can be further adjusted according to specific conditions. Any simple improvement of the method of the present invention under the premise of the concept of the present invention belongs to the scope of protection required by the present invention.
[0033] Please refer to Figures 1-4 , the present invention provides a multi-athlete tracking, identification and timing system based on a smart phone, including a global detector 1 and a tracker and SORT algorithm 11. The global detector 1 is connected to an input image preprocessing module 2, the global detector 1 is connected to a target detection model module 3, the global detector 1 is connected to a class recognition and localization module 4, the global detector 1 is connected to a post-processing and filtering module 5, the global detector 1 is connected to an athlete identification module 6, the global detector 1 is connected to a non-athlete identification module 7, the global detector 1 is connected to a finish line identification module 8, the finish line identification module 8 is connected to a finish line feature module 9, the finish line identification module 8 is connected to a position detection module 10, the tracker and SORT algorithm 11 is connected to an athlete tracking module 12, the tracker and SORT algorithm 11 is connected to a dynamic tracking module 13 for athletes and non-athletes, and the tracker and SORT algorithm 11 is connected to a finish line detection and position tracking module 14.
[0034] The main function of the global detector 1 is to identify and classify different targets in the picture, such as athletes, non-athletes and the finish line.
[0035] The input image preprocessing module 2 can preprocess the image. Operations such as scaling, normalization and color space conversion can improve the detection efficiency and accuracy of multi-athlete timing and tracking.
[0036] The target detection model module 3 can process the image using a deep learning-based target detection algorithm and can also detect potential targets from it.
[0037] The class recognition and localization module 4 can classify each athlete, non-athlete and finish line and also provide its bounding box to indicate the position of the target in the image.
[0038] The post - processing and filtering module 5 can handle non - maximum suppression and can also remove overlapping bounding boxes to ensure that each target has only one most accurate detection result.
[0039] In this embodiment, by setting the global detector 1, the input image pre - processing module 2, the post - processing and filtering module 5, and the tracker and SORT algorithm 11, the global detector 1 can receive the pre - processed image or video frame after passing through the input image pre - processing module 2. Through operations such as scaling, normalization, and color - space conversion, the detection efficiency and accuracy of multi - athlete timing tracking can be improved. The global detector 1 can cooperate with the post - processing and filtering module 5 to handle non - maximum suppression and remove overlapping bounding boxes to ensure that each target has only one most accurate detection result, thereby ensuring the accuracy of the multi - athlete timing tracking detection result. The tracker and SORT algorithm 11 can track each athlete in consecutive video frames through the SORT algorithm system. When tracking multiple targets, the tracker and SORT algorithm 11 can effectively handle the dynamic changes of multiple targets in the scene, including the interaction between athletes and non - athletes, thereby avoiding the situation of confusing or losing targets. And through cooperation with the global detector 1, the tracker and SORT algorithm 11 can be used to track the change of the finish line in real - time. In scenarios such as track races, the finish line usually appears at a specific frame position. The tracker and SORT algorithm 11 can track the exact position of the finish line and can help athletes with position correction, so as to ensure that each athlete is uniquely identified and tracked.
[0040] In this embodiment, by setting the target detection model module 3 and the category recognition and localization module 4, the target detection model module 3 can use a deep - learning - based target detection algorithm to process the image and can detect potential targets from it. At the same time, in cooperation with the global detector 1, it can automatically learn the features of various targets in the image through a convolutional neural network to efficiently identify and locate athletes, non - athletes, and the finish line in the picture. The category recognition and localization module 4 can classify each athlete, non - athlete, and the finish line and also provide their bounding boxes to indicate the position of the target in the image, thereby ensuring the detection accuracy of each athlete, non - athlete, and the finish line.
[0041] A multi - athlete tracking, recognition, and timing method based on a smartphone, applied to the multi - athlete tracking, recognition, and timing system based on a smartphone described in the above - mentioned embodiment, includes the following steps:
[0042] S1. The application software is started;
[0043] S2. The mobile phone camera is called;
[0044] S3. A frame of picture of the video collected by the mobile phone is obtained;
[0045] S4, First-level detector;
[0046] S5, Tracker;
[0047] S6, Bib detector;
[0048] S7, Bib character detector;
[0049] S8, Finish-line action determination detector;
[0050] S9, Data integration, forming data labels;
[0051] S10, Record the moment when the athlete crosses the finish line;
[0052] S11, Write the data and pictures into the database.
[0053] The working principle and usage process of the present invention: The referee holds a mobile phone to take pictures of the athletes, and real-time collects the videos before and after the moment when the athletes cross the finish line. The global detector 1 will detect the movement process of the athletes in real time. The input image preprocessing module 2 will perform operations such as scaling, normalization, and color space conversion to preprocess the image so as to improve the detection efficiency and accuracy of multi-athlete timing and tracking. The target detection model module 3 will process the image and detect potential targets from it in order to efficiently identify and locate the athletes, non-athletes, and finish line in the picture. The category recognition and localization module 4 will classify the athletes, non-athletes, and finish line, and will provide bounding boxes to indicate the positions of the targets in the image. The post-processing and filtering module 5 will perform post-processing on the detected images to ensure that there is only one most accurate detection result for each target. The tracker and SORT algorithm 11 will track the athletes, non-athletes, and finish line in real time, and will ensure that each athlete is uniquely identified and tracked, and will also help the athletes correct their positions. By tracking and identifying the athletes, identifying the finish line, and determining the moment when the athletes cross the finish line, the finish time of the athletes can be obtained, and the finish time of the athletes will be generated on the mobile phone and the athlete score table will be directly generated, so that the timing function of the athletes can be accurately completed. At the same time, the images and videos related to the athletes will be recorded and saved, and the referee can move the mobile phone during the process of holding the mobile phone, thereby being able to avoid being blocked by other athletes or staff.
[0054] Although the embodiments of the present invention have been shown and described, for those of ordinary skill in the art, it can be understood that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirits of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.
Claims
1. A multi-athlete tracking, identification and timing system based on a smartphone, comprising a global detector (1) and a tracker and a SORT algorithm (11), characterized in that: The global detector (1) is connected to an input image preprocessing module (2), the global detector (1) is connected to a target detection model module (3), the global detector (1) is connected to a category recognition and positioning module (4), the global detector (1) is connected to a post-processing and filtering module (5), the global detector (1) is connected to an athlete recognition module (6), the global detector (1) is connected to a non-athlete recognition module (7), the global detector (1) is connected to a finish line recognition module (8), the finish line recognition module (8) is connected to a finish line feature module (9), the finish line recognition module (8) is connected to a position detection module (10), the tracker and the SORT algorithm (11) are connected to an athlete tracking module (12), the tracker and the SORT algorithm (11) are connected to an athlete and non-athlete dynamic tracking module (13), and the tracker and the SORT algorithm (11) are connected to a finish line detection and position tracking module (14).
2. The multi-athlete tracking, identification and timing system based on a smart phone according to claim 1, characterized in that: The main function of the global detector (1) is to identify and classify different objects in the picture, such as athletes, non-athletes and the finish line.
3. The multi-athlete tracking, identification and timing system based on a smart phone according to claim 1, characterized in that: The input image preprocessing module (2) is capable of preprocessing the image.
4. The multi-athlete tracking, identification and timing system based on a smart phone according to claim 1, characterized in that: The target detection model module (3) can process images using a deep learning-based target detection algorithm and can also detect potential targets therefrom.
5. The multi-athlete tracking, identification and timing system based on a smart phone according to claim 1, characterized in that: The category recognition and localization module (4) is capable of classifying each athlete, non-athlete and finish line, and also provides its bounding box to indicate the location of the object in the image.
6. The smart phone-based multi-athlete tracking, identification and timing system according to claim 1, characterized in that: The post-processing and filtering module (5) can handle non-maximum suppression and remove overlapping bounding boxes to ensure that each target has only one most accurate detection result.
7. A method for tracking, identifying and timing multiple athletes based on a smart phone, applied to a system for tracking, identifying and timing multiple athletes based on a smart phone as claimed in any one of claims 1 to 6, characterized in that: The following steps are involved: S1, application software starts; S2, call the mobile phone camera; S3, obtaining a frame of the video captured by the mobile phone; S4, first stage detector; S5, tracker; S6, bib number detector; S7, number cloth character detector; S8, finishing line action determination detector; S9, data integration, forming data labels; S10, record the moment when the athlete crosses the finish line; S11. Write data and pictures into the database.