Real-time video mark display method for athletes in running competition and related equipment
Through the global detection and real -time tracking of the running game video, combined with number cloth detection and data label generation, the problem of complex and inaccurate in the existing technology of the existing technology, and the accurate recognition of the athlete's identity and the time of the lines of the line. And showing the accuracy of referee punishment.
Patent Information
- Application Number
- CN202411410422.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-10-10
- Publication Date
- 2025-01-28
AI Technical Summary
The prior art is complicated and inaccurate in determining the timing of athletes rushing through the line, especially due to the distortion or obstruction of the number cloth, which makes it difficult to determine the athlete's identity, which affects the accuracy of the referee's penalty.
By obtaining the original video of the camera of the end line, the global detection and identification athlete, non -athletes and end line lines are performed in real time, tracking the athlete trajectory in real time, determining the line action and time, detecting and dividing the number cloth image in the video. Output athlete's participation number, and bind the number, line time, and participation number to generate data labels, combine data tags with original video, forming new videos with data labels.
The accurate judgment and identity recognition of the athletes of the athletes were realized. By showing the athlete's participation number and rushing time on the large screen on the competition site, the accuracy of the referee was improved.
Smart Images
Figure CN119383421A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of video data processing technology, and in particular to a method, device, system, electronic device and computer-readable storage medium for displaying real-time video markings of athletes in a running competition. Background Art
[0002] At present, running competitions such as marathons mainly use manual methods to determine the time of crossing the finish line, or use the method of regional demarcation to determine the time of crossing the finish line. The determination process is complicated or the determination results are inaccurate. In addition, the athlete's movement amplitude is large at the moment of crossing the finish line. The number cloth may be distorted or blocked, resulting in the inability to effectively identify the number cloth at the time of crossing the finish line, and then the athlete's identity cannot be accurately determined, which affects the on-site referee's penalty. Summary of the invention
[0003] The technical problem to be solved by the present invention is to provide a method, device, system, electronic device and computer-readable storage medium for displaying real-time video markings of athletes in a running competition.
[0004] In order to solve the above technical problems, the present invention adopts the following technical solutions:
[0005] A real-time video marking display method for athletes in a running competition comprises the following steps: obtaining an original video shot by a camera facing a finish line; performing global detection on the entire video screen of the original video to identify athletes, non-athletes and the finish line in the screen; performing real-time tracking on the identified athletes and finish line to obtain the trajectory of the athletes and give the athletes a unique number; determining the athlete's finishing action to obtain the athlete's finishing time; detecting and segmenting a number cloth image in the athlete's trajectory image; detecting and segmenting a competition number image on the number cloth in the number cloth image, and outputting the athlete's competition number in a continuous text state; binding the athlete's number, finishing time and competition number to form a corresponding relationship between the competition number and finishing time, and generating a data label in a graphical form; fusing the data label with the original video to form a new video; and outputting the new video with the data label.
[0006] A real-time video mark display device for athletes in a running competition comprises: a video acquisition module for acquiring an original video shot by a camera facing a finish line; a global detector for performing global detection on the entire video screen of the original video, identifying athletes, non-athletes and a finish line in the screen; a tracker for real-time tracking of the identified athletes and finish lines, obtaining the track of the athletes, and giving the athletes a unique number; a finish line action determination detector for determining the finish line action of the athletes and obtaining the finish line time of the athletes; a number cloth detector for detecting and segmenting a number cloth image in the athlete's track image; a number cloth character detector for detecting and segmenting a competition number image on the number cloth in the number cloth image, and outputting the athlete's competition number in a continuous text state; a data label generation module for binding the athlete's number, finish line time and competition number to form a corresponding relationship between the competition number and the finish line time, and generating a data label in a graphical form; a video fusion module for fusing the data label with the original video to form a new video; and a video output module for outputting the new video with the data label.
[0007] A system for displaying real-time video marks of athletes in a running competition, comprising at least one camera facing the finish line, a display device installed at the competition site, and a display device for displaying real-time video marks of athletes in a running competition, wherein the camera is used to continuously track and shoot athletes, and the display device is used to receive new videos with data tags output by the real-time video mark display device for athletes in a running competition and play and output the new videos; the real-time video mark display device for athletes in a running competition comprises: a video acquisition module, used to acquire original videos shot by the camera facing the finish line; a global detector, used to perform global detection on the entire video screen of the original video, and identify athletes, non-athletes and the finish line in the screen; a tracker, used to track the identified athletes and the finish line The system includes a control module, a control module for detecting and segmenting the image of the number cloth in the image of the athlete's trajectory; a control module for detecting and segmenting the image of the competition number on the number cloth in the image of the athlete's trajectory, and outputting the athlete's competition number in a continuous text state; a data label generation module, a control module for binding the athlete's number, the time of crossing the finish line, and the competition number to form a corresponding relationship between the competition number and the time of crossing the finish line, and generating a data label in a graphical form; a video fusion module, a control module for merging the data label with the original video to form a new video; and a video output module, a control module for outputting a new video with a data label.
[0008] An electronic device includes at least one processor and at least one memory in communication with the processor, wherein the memory stores program instructions, and when the program instructions are called and executed by the processor, the above-mentioned method for displaying real-time video markings of athletes in a running competition is implemented.
[0009] A computer-readable storage medium stores program instructions, which, when executed by a processor, implement the above-mentioned method for displaying real-time video markings of athletes in a running competition.
[0010] The beneficial technical effect of the present invention is that the present invention can effectively determine the time and identity of the athlete crossing the finish line by continuously tracking the athlete before and after the athlete crosses the finish line, display the real-time video of the athlete crossing the finish line on the large screen at the competition site, and mark the athlete's competition number and finishing time on each athlete in real time, providing a reference for the referee's judgment on site, which can greatly improve the accuracy of the referee's judgment. BRIEF DESCRIPTION OF THE DRAWINGS
[0011] Figure 1 It is a schematic flow chart of a method for displaying real-time video markings of athletes in a running competition according to the present invention;
[0012] Figure 2 This is a functional structure diagram of the real-time video marking display device for athletes in a running competition of the present invention;
[0013] Figure 3 The present invention is a structural schematic diagram of a system for displaying real-time video markings of athletes in a running competition. DETAILED DESCRIPTION
[0014] In order to enable those skilled in the art to more clearly understand the objectives, technical solutions and advantages of the present invention, the present invention is further described below in conjunction with the accompanying drawings and embodiments.
[0015] The invention provides a method for displaying real-time video markings of athletes in a running competition.
[0016] like Figure 1 As shown, in one embodiment of the present invention, the method for displaying real-time video markers of athletes in a running competition includes steps S10 to S90:
[0017] S10: Obtain the original video shot by the camera facing the finish line.
[0018] The original video obtained in this step is a real-time video stream shot by a camera facing the finish line, which can clearly show the finish line of the race.
[0019] S20: Perform global detection on the entire video screen of the original video to identify athletes, non-athletes and a finish line in the screen.
[0020] In specific implementation, a first-order or second-order method is used to perform global detection on the entire video screen of the original video to identify athletes, non-athletes, and the finish line in the screen. The first-order method directly regresses the position and category of the target in the output layer, without the need to generate candidate regions. Typical representatives of this type of algorithm include the YOLO (You Only Look Once, an algorithm that uses convolutional neural networks for target detection) algorithm and the SSD (Single Shot MultiBox Detector, an algorithm for target detection based on deep learning) algorithm. The second-order method first generates a series of candidate regions (RegionProposals), and then classifies and refines the positions of these candidate regions. Typical representatives of this type of algorithm include the R-CNN series of algorithms based on regionproposal, such as R-CNN, FastR-CNN, FasterR-CNN and other algorithms.
[0021] S30: Track the identified athletes and finish line in real time, obtain the athlete's trajectory, and give the athlete a unique number.
[0022] In the specific implementation, the SORT (Simple Online and Realtime Tracking) algorithm is used to track the identified athletes and the finish line in real time to obtain the athlete's trajectory.
[0023] The SORT algorithm includes three steps: target detection, Kalman filtering and data association.
[0024] Object detection: First, use the object detection algorithm (such as YOLO, SSD or Faster R-CNN) to obtain the object position information in the current frame, including the object's bounding box, confidence, etc. This step is the basis of the SORT algorithm, which provides the initial position information of the object in each frame.
[0025] Kalman filter: The SORT algorithm uses the Kalman filter to model and predict the motion of the target. The Kalman filter combines the current state of the target with the motion model to estimate the position and velocity of the target in the next frame. This step is the key to the SORT algorithm to achieve target prediction. It predicts the possible position of the target in the next frame based on the target's motion pattern.
[0026] Data association: The SORT algorithm introduces a data association method based on the Hungarian algorithm to establish an association relationship between the current frame and the previous frame. By calculating the distance and speed difference between the targets, the targets in adjacent frames are matched to achieve continuous tracking of the targets. This step is the core of the SORT algorithm to achieve target tracking. It matches the detected target with the previously tracked target through the data association method to form the target's trajectory.
[0027] S40, judging the athlete's finishing action and obtaining the athlete's finishing time.
[0028] In specific implementation, algorithms such as ResNet can be used to determine the athlete's finishing action, thereby obtaining the athlete's finishing time.
[0029] S50: Detect and segment the number cloth image from the athlete's track image.
[0030] In specific implementation, an end-to-end deep learning machine vision detection method, such as the YOLO algorithm, is used to detect and segment the number cloth image in the athlete's trajectory image.
[0031] S60: Detect and segment the competition number image on the number cloth from the number cloth image, and output the athlete's competition number in a continuous text state.
[0032] In specific implementation, an end-to-end deep learning machine vision detection method is used, such as the YOLO algorithm, to detect and segment the competition number image on the number cloth in the number cloth image.
[0033] S70, binding the athlete's serial number, finishing time, and competition number to form a corresponding relationship between the competition number and the finishing time, and generating a data label in a graphical form.
[0034] For an athlete with a unique number, the athlete's number, finishing time, and competition number are bound together, and finally a corresponding relationship between the competition number and the finishing time is formed, and a data label in a graphical form is generated. According to this method, the number, finishing time, and competition number of each athlete are bound together, and a corresponding data label is generated.
[0035] S80: Merge the data label with the original video to form a new video.
[0036] S90, output a new video with data labels.
[0037] The new video with data labels is output to the display or large screen at the competition site for playback and display, and the athlete's competition number and finishing time are marked on each athlete in real time, providing a reference for the referees on site, which can greatly improve the accuracy of the referee's judgment.
[0038] The present invention also provides a real-time video mark display device for athletes in a running competition.
[0039] like Figure 2 As shown, in one embodiment of the present invention, the real-time video mark display device for running competition athletes includes:
[0040] The video acquisition module 10 is used to acquire the original video shot by the camera facing the finish line. The original video acquired by the video acquisition module 10 is a real-time video stream shot by the camera facing the finish line, which can clearly show the finish line of the race.
[0041] The global detector 20 is used to perform global detection on the entire video screen of the original video, and identify the athletes, non-athletes and finish line in the screen. The global detector 20 uses a first-order or second-order method to perform global detection on the entire video screen of the original video, and identify the athletes, non-athletes and finish line in the screen. The first-order method directly regresses the position and category of the target in the output layer, and does not need to generate the candidate region stage. Typical representatives of this type of algorithm include the YOLO (You Only Look Once, an algorithm that uses a convolutional neural network for target detection) algorithm and the SSD (Single Shot Multi Box Detector, a target detection algorithm based on deep learning) algorithm. The second-order method first generates a series of candidate regions (Region Proposals), and then classifies and refines the positions of these candidate regions. Typical representatives of this type of algorithm include the R-CNN series of algorithms based on region proposal, such as R-CNN, FastR-CNN, FasterR-CNN and other algorithms.
[0042] The tracker 30 is used to track the identified athletes and the finish line in real time, obtain the athlete's trajectory, and give the athlete a unique number. The tracker 30 uses the SORT (Simple Online and Realtime Tracking) algorithm to track the identified athletes and the finish line in real time to obtain the athlete's trajectory.
[0043] The finish line action determination detector 40 is used to determine the finish line action of the athlete and obtain the finish line time of the athlete. The finish line action determination detector 40 can use algorithms such as ResNet to determine the finish line action of the athlete, thereby obtaining the finish line time of the athlete.
[0044] The number cloth detector 50 is used to detect and segment the number cloth image in the athlete's track image. The number cloth detector 50 uses an end-to-end deep learning machine vision detection method, such as the YOLO algorithm, to detect and segment the number cloth image in the athlete's track image.
[0045] The number cloth character detector 60 is used to detect and segment the competition number image on the number cloth in the number cloth image, and output the athlete's competition number in a continuous text state. The number cloth character detector 60 uses an end-to-end deep learning machine vision detection method, such as the YOLO algorithm, to detect and segment the competition number image on the number cloth in the number cloth image.
[0046] The data label generation module 70 is used to bind the athlete's serial number, finishing time, and competition number to form a corresponding relationship between the competition number and the finishing time, and generate a data label in a graphical form.
[0047] The video fusion module 80 is used to fuse the data label with the original video to form a new video.
[0048] The video output module 90 outputs the new video with the data label. The video output module 90 outputs the new video with the data label to the display or large screen at the competition site for playback and display, and marks the athlete's competition number and finishing time on each athlete in real time, providing a reference for the referee's judgment on the spot, which can greatly improve the accuracy of the referee's judgment.
[0049] The invention also provides a real-time video marking display system for athletes in a running competition.
[0050] like Figure 3 As shown, in one embodiment of the present invention, the running competition athlete real-time video mark display system includes at least one camera 101 facing the finish line, a display device (display or large screen) 200 installed at the competition site, and a running competition athlete real-time video mark display device 100, wherein the camera 101 is used to continuously track and shoot the athlete, and the running competition athlete real-time video mark display device 100 adopts the following method: Figure 2 In the real-time video marking display device for running competition athletes in the illustrated embodiment, the display device 200 is used to receive a new video with a data tag output by the real-time video marking display device 100 for running competition athletes and to play and output the new video, that is, it is used to receive a new video with a data tag output by the video output module of the real-time video marking display device 100 for running competition athletes and to play and output the new video.
[0051] The present invention also provides an electronic device, comprising a processor and a memory in communication with the processor, wherein the memory is used to store various types of data to support operations on the device, and the data may include instructions for any application or method used to operate on the device, as well as data related to the application. The memory may be implemented by one or more volatile or non-volatile storage devices of any type or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk. The processor may be one or more, and the processor is used to control the overall operation of the device, and completes the operation by calling and executing the program instructions stored in the memory. Figure 1 The various steps of the method for displaying real-time video markings of athletes in a running competition in the illustrated embodiment.
[0052] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores program instructions, and when the program instructions are executed by a processor, the steps of the above-mentioned method for displaying real-time video marks of athletes in a running competition are implemented. The storage medium may be the above-mentioned memory including the program instructions, and the above-mentioned program instructions may be executed by the processor to complete Figure 1 The various steps of the method for displaying real-time video markings of athletes in a running competition in the illustrated embodiment.
[0053] The above description is only a preferred embodiment of the present invention, and does not limit the present invention in any form. Those skilled in the art can make various equivalent changes and improvements based on the above embodiments, and all equivalent changes or modifications made within the scope of the claims should fall within the protection scope of the present invention.
Claims
1. A method for displaying real-time video markings of athletes in a running competition, characterized in that: The steps include: Get the original video shot by the camera facing the finish line; Performing global detection on the entire video screen of the original video to identify athletes, non-athletes and a finish line in the screen; Track the identified athletes and finish lines in real time, obtain the athlete's trajectory, and give the athlete a unique number; Determine the athlete's finishing action and obtain the athlete's finishing time; Detect and segment the bib image in the athlete's track image; Detect and segment the race number image on the race number cloth in the race number cloth image, and output the athlete's race number in a continuous text state; Bind the athlete's number, finishing time, and race number to form a corresponding relationship between the race number and the finishing time, and generate a data label in a graphical form; Merge the data labels with the original video to form a new video; Output a new video with data labels.
2. The method for displaying real-time video marks of athletes in a running competition as claimed in claim 1, characterized in that: The YOLO algorithm, the SSD algorithm, the R-CNN algorithm, the Fast R-CNN algorithm or the Faster R-CNN algorithm is used to perform global detection on the entire video screen of the original video to identify athletes, non-athletes and the finish line in the screen.
3. The method for displaying real-time video marks of athletes in a running competition as claimed in claim 1, characterized in that: The SORT algorithm is used to track the identified athletes and finish line in real time to obtain the athlete’s trajectory.
4. The method for displaying real-time video marks of athletes in a running competition as claimed in claim 1, characterized in that: The ResNet algorithm is used to determine the athlete's finishing action and obtain the athlete's finishing time.
5. The method for displaying real-time video marks of athletes in a running competition as claimed in claim 1, characterized in that: An end-to-end deep learning machine vision detection method is used to detect and segment the number cloth images in the athlete's trajectory images.
6. The method for displaying real-time video marks of athletes in a running competition as claimed in claim 1, characterized in that: An end-to-end deep learning machine vision detection method is used to detect and segment the race number image on the number cloth in the number cloth image.
7. A real-time video mark display device for athletes in a running competition, characterized in that: Included are: A video acquisition module is used to acquire the original video shot by a camera facing the finish line; A global detector, used for performing global detection on the entire video screen of the original video to identify athletes, non-athletes and a finish line in the screen; Tracker, used to track the identified athletes and finish line in real time, obtain the athlete's trajectory, and give the athlete a unique number; A finishing action determination detector is used to determine the finishing action of the athlete and obtain the finishing time of the athlete; A bib number detector, used to detect and segment the bib number image in the athlete's track image; The number cloth character detector is used to detect and segment the competition number image on the number cloth in the number cloth image, and output the athlete's competition number in a continuous text state; The data label generation module is used to bind the athlete's number, finishing time, and competition number to form a corresponding relationship between the competition number and the finishing time, and generate a data label in a graphical form; Video fusion module, used to fuse data labels with original videos to form new videos; Video output module, outputs new videos with data labels.
8. A real-time video marking display system for athletes in a running competition, characterized in that: The system comprises at least one camera facing the finish line, a display device installed at the competition site, and a real-time video mark display device for athletes in a running competition, wherein the camera is used to continuously track and shoot the athletes, and the display device is used to receive a new video with a data tag output by the real-time video mark display device for athletes in a running competition and play and output the new video; The real-time video marking display device for running competition athletes includes: A video acquisition module is used to acquire the original video shot by a camera facing the finish line; A global detector, used for performing global detection on the entire video screen of the original video to identify athletes, non-athletes and a finish line in the screen; Tracker, used to track the identified athletes and finish line in real time, obtain the athlete's trajectory, and give the athlete a unique number; A finishing action determination detector is used to determine the finishing action of the athlete and obtain the finishing time of the athlete; A bib number detector, used to detect and segment the bib number image in the athlete's track image; The number cloth character detector is used to detect and segment the competition number image on the number cloth in the number cloth image, and output the athlete's competition number in a continuous text state; The data label generation module is used to bind the athlete's number, finishing time, and competition number to form a corresponding relationship between the competition number and the finishing time, and generate a data label in a graphical form; Video fusion module, used to fuse data labels with original videos to form new videos; Video output module, outputs new videos with data labels.
9. An electronic device, characterized in that: The electronic device includes at least one processor and at least one memory in communication with the processor, wherein the memory stores program instructions, and when the program instructions are called and executed by the processor, the real-time video marking display method for running competition athletes described in any one of claims 1 to 6 is implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores program instructions, and when the program instructions are executed by a processor, the method for displaying real-time video markings of athletes in a running competition according to any one of claims 1 to 6 is implemented.