Billiard game intelligent management system based on AI visual analysis

Abstract

The invention relates to a billiard game intelligent management system based on AI visual analysis, and belongs to the technical field of management systems, and the system comprises a high-definition camera which is installed above a billiard table and is used for collecting a game video stream; the AI edge calculation box is internally provided with a GPU / NPU, runs a deep learning model and is used for performing real-time analysis on the competition video stream acquired by the high-definition camera; the billiard recognition model is a ball detection and trajectory tracking model trained based on a YOLO or DeepLab algorithm; the rule engine module is used for carrying out logic judgment according to billiard match rules; the mobile application / large screen system is used for displaying real-time scores, wonderful playback and match data; and the storage server is used for storing the video clips, the match logs and the wonderful instant marks.

Description

Technical Field

[0001] This invention relates to the field of management system technology, and in particular to an intelligent management system for billiards competitions based on AI visual analysis. Background Technology

[0002] With the development of artificial intelligence and computer vision technology, intelligent management of sports events is gradually becoming a trend. Billiards, as a widely popular indoor sport, has characteristics such as clear rules, standardized movements, and fixed space, making it suitable for AI visual recognition.

[0003] Currently, some high-end billiard halls or tournaments have begun to try introducing video monitoring and playback functions, but most still rely on manual operation, which has the following problems: 1. Non-intelligent: It only provides video recording function and cannot automatically identify key events such as goals and fouls; 2. Low playback efficiency: Users need to manually drag the timeline to find highlights, resulting in a poor user experience; 3. Scoring relies on manual intervention: Scoring in the competition is done manually by players or referees, which is prone to errors; 4. Lack of interactivity: Viewers cannot access match data and highlight replays in real time; 5. System independence: Most existing AI analysis systems are general-purpose platforms and have not been deeply optimized for billiards rules.

[0004] In the existing technology, there are some sports motion recognition systems based on cameras (such as AI analysis for table tennis and badminton), but their recognition logic is quite different from that of billiards, and cannot be directly transferred and applied. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art, the present invention provides a Doppler flow detection device for a billiards competition intelligent management system based on AI visual analysis. This device is easy to operate, low in cost, and can keep the detection surface of the detection device and the water surface to be detected at the optimal detection angle.

[0006] To achieve the above-mentioned objectives, this invention provides an intelligent management system for billiards matches based on AI visual analysis, comprising: High-definition camera: Installed above the billiard table to capture video streams of the game; AI edge computing box: Built-in GPU / NPU, running deep learning models for real-time analysis of the competition video stream captured by the high-definition camera; Billiards recognition model: a ball detection and trajectory tracking model trained based on YOLO or DeepLab algorithms; Rules engine module: Performs logical judgments based on the rules of billiards competition; Mobile application / large screen system: used to display real-time scores, highlights, and match data; Storage server: Used to store video clips, match logs, and highlight markers.

[0007] Furthermore, the AI ​​edge computing box receives the RTSP / H.264 video stream from the high-definition camera and processes it through the following steps: Video preprocessing: Denoising, distortion correction, and illumination compensation are performed on the video stream images captured by the high-definition camera; Sphere detection: Using a CNN model to identify the positions of 15 target spheres and the white sphere; Trajectory tracking: Tracking the trajectory of a sphere using SORT or DeepSORT algorithms; Pocket entry recognition: Combines pocket opening area detection with ball disappearance logic to determine a goal event; Foul detection: Identifies actions such as the cue ball being pocketed, not touching one's own ball, and jump ball fouls; Spectacular Shot Judgment: Scoring is based on three dimensions: difficulty of scoring, number of consecutive hits, and positioning accuracy, and the shot is marked as a spectacular moment.

[0008] Furthermore, the workflow of the AI-based visual analysis-based intelligent management system for billiards matches is as follows: Intelligent scoring and refereeing: After the AI ​​edge computing box identifies a goal, it automatically updates the current player's score. After a foul is determined, the ball possession is automatically switched, and the type of foul is indicated. Real-time score synchronization to mobile applications / large screen systems enables intelligent referee functionality; Highlight Recognition and Replay: When the system detects events such as clearing the table, potting a long ball, or solving a snooker, it automatically marks the timestamp. Users can quickly view the highlights recommended by AI by clicking the "Highlights" button in the mobile app; It supports adding special effects to enhance the viewing experience, such as slow motion, spotlights, and sound effects; Match statistics: Automatically generate match reports, including: number of goals, success rate, number of fouls, average shot time, etc. Supports historical match review and data comparison analysis.

[0009] Compared with the prior art, the beneficial effects of the present invention are: Highly automated: Scoring, refereeing, and replay marking can be achieved without human intervention, reducing operating costs; High recognition accuracy: The AI ​​model optimized for billiards scenarios achieves a pocket recognition accuracy of over 95%; High real-time performance: The edge computing box processes locally with a latency of less than 200ms, meeting the requirements of real-time interaction; Excellent user experience: Supports automatic push notifications of highlights, one-click playback, and special effects playback to enhance the viewing experience; Highly scalable: Supports multiple billiards rules (8-ball, 9-ball, snooker) and can be adapted to different tournament scenarios; High data value: Accumulated competition data can be used for player rating, training analysis, and business operations. Attached Figure Description

[0010] Figure 1 This is a structural block diagram of an embodiment of the present invention; Figure 2 This is a flowchart illustrating the processing flow of the AI ​​edge computing box of the present invention. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the invention, but not all embodiments. The embodiments of the present invention are described below with reference to the accompanying drawings.

[0012] The shortcomings of current technology in billiards competitions are: Traditional billiards matches rely on manual scoring, which is prone to misjudgments and omissions. The video surveillance system only records video and does not have automatic event recognition capabilities; Users cannot easily access exciting moments such as "goal", "clearance", and "snooker"; The lack of a real-time scoreboard and intelligent referee function affects the professionalism and entertainment value of the game; Existing AI systems do not model the physical rules of billiards (such as ball order and foul judgment), resulting in low recognition accuracy.

[0013] The purpose of this invention is to provide an intelligent management system for billiards matches based on an AI edge computing box. By deploying a camera above the billiard table and connecting it to an AI edge computing box, the system utilizes a trained deep learning model to analyze the video stream in real time, automatically identifying ball trajectories, potting events, and fouls, and generating a scoreboard, highlight markers, and automatic replay clips. The system can be integrated with a mobile application or large-screen system to achieve functions such as intelligent refereeing, highlight shot effects playback, and match data statistics, thereby improving the automation level and user experience of billiards matches. like Figure 1 As shown, a smart management system for billiards matches based on AI visual analysis includes: High-definition cameras are installed above the billiard table during billiards matches to capture video streams of the game. These cameras have a resolution of 720p or higher and can capture high-quality video images of the match. The AI ​​edge computing box has a built-in GPU / NPU to run deep learning models for real-time analysis of the competition video stream captured by high-definition cameras. The AI ​​edge computing box has a built-in GPU or NPU. GPUs were originally designed to accelerate graphics rendering and are good at handling large-scale parallel tasks (such as 3D graphics computing); NPUs are dedicated chips designed for artificial intelligence (especially neural network operations). Their core advantage is that they can efficiently perform core AI operations such as matrix multiplication and convolution. Specifically, you can choose to build an NPU into the AI ​​edge computing box. Billiard ball recognition model: A ball detection and trajectory tracking model trained using YOLO or DeepLab algorithms; Rules engine module: Performs logical judgments based on the rules of billiards competition; Mobile applications / large screen systems are used to display real-time scores, highlights, and match data. Both mobile applications and large screen systems can be used to display real-time scores, highlights, and match data. Mobile applications can be installed on mobile devices for convenient viewing by users on their mobile devices. Storage server: Used to store video clips, match logs, and highlight markers.

[0014] The intelligent management system for billiards competitions proposed in this application combines high-definition cameras, AI edge computing boxes, deep learning models, real-time video stream processing, and mobile application technologies. It primarily addresses the problems of traditional billiards competitions, such as reliance on manual scoring, lack of automatic replay and recording of highlights, and poor audience interactivity. By deploying AI boxes to collect and intelligently analyze real-time video of billiards matches, it enables functions such as automatically recognizing shots, judging fouls, generating highlights, intelligent scoring, and replay, thereby enhancing the intelligence, automation, and entertainment value of billiards competitions and promoting the digital upgrade of the sport.

[0015] refer to Figure 2 The AI ​​edge computing box receives RTSP / H.264 video streams from a high-definition camera and processes them through the following steps: Video preprocessing: Denoising, distortion correction, and illumination compensation are performed on the video stream images captured by the high-definition camera; Sphere detection can be performed by using a CNN model to identify the positions of 15 target spheres and the white sphere, or by using a YOLOv8 model to identify the positions of 15 target spheres and the white sphere. Trajectory tracking: Tracking the trajectory of a sphere using SORT or DeepSORT algorithms; specifically, the SORT algorithm primarily relies on motion information for tracking. It predicts the target's position in the next frame using Kalman filtering, and then uses the Hungarian algorithm (based on IoU distance) to match the predicted bounding box with the detection box, thereby achieving target tracking. Its advantage is its high speed (over 200 frames / second).

[0016] The DeepSORT algorithm (Deep Learning for Object Tracking with SORT) is an upgraded version of SORT. It extracts additional appearance features of the target (such as color and texture) on the basis of SORT and combines appearance features with motion information (IoU) to build a more comprehensive similarity measure, thereby significantly reducing the frequency of ID switching (by 45%). This makes DeepSORT more stable in complex scenes (such as dense pedestrians and occlusion). The system identifies balls entering the pocket by combining the detection of the pocket opening area with the logic of the ball disappearing. By combining the detection of the pocket opening area with the case of the ball disappearing, the system can determine whether the ball has entered the pocket. Foul detection: Identify actions such as "cup ball pocketed", "not touching one's own ball", and "jump shot foul", and complete the foul detection; The criteria for judging a highlight shot are based on three dimensions: the difficulty of scoring, the number of consecutive hits, and the accuracy of positioning, marking the highlight moments.

[0017] The workflow of the AI-based visual analysis-based intelligent management system for billiards matches is as follows: Intelligent scoring and refereeing: The AI ​​edge computing box identifies the ball that goes into the pocket, determines the player who scored based on the ball after it enters the pocket, and then automatically updates the current player's score, thus realizing the function of intelligent score calculation. After a foul is detected, the right to strike the ball is automatically switched, and the type of foul is indicated. The system determines the player's foul based on the foul detection. After the current player commits a foul, the right to strike the ball is automatically switched to the opponent, and the specific type of foul can be indicated on the large screen system. The scores are synchronized in real time to mobile applications and large-screen systems to enable intelligent refereeing. Highlight Recognition and Replay: When the system detects events such as "clearing the table", "long-range shot", or "snooker escape", it automatically marks the timestamp. Users can quickly view AI-recommended highlight moments by clicking the "Highlights" button in the mobile app; It supports adding special effects to enhance the viewing experience, such as slow motion, spotlights, and sound effects; Match statistics: Automatically generate match reports, including: number of goals, success rate, number of fouls, average shot time, etc. Supports historical match review and data comparison analysis.

[0018] Compared with the prior art, the present invention has the following advantages: Highly automated: Scoring, refereeing, and replay marking can be achieved without human intervention, reducing operating costs; High recognition accuracy: The AI ​​model optimized for billiards scenarios achieves a pocket recognition accuracy of over 95%; High real-time performance: The edge computing box processes locally with a latency of less than 200ms, meeting the requirements of real-time interaction; Excellent user experience: Supports automatic push notifications of highlights, one-click playback, and special effects playback to enhance the viewing experience; Highly scalable: Supports multiple billiards rules (8-ball, 9-ball, snooker) and can be adapted to different tournament scenarios; High data value: Accumulated competition data can be used for player rating, training analysis, and business operations.

[0019] The technical solution of the present invention has been described above in conjunction with specific embodiments. However, it should be noted that the above descriptions are only for explaining the solution of the present invention and should not be construed as a specific limitation on the scope of protection of the invention in any way. Based on this explanation, those skilled in the art can conceive of other specific embodiments or equivalent substitutions of the present invention without creative effort, and all such embodiments or substitutions will fall within the scope of protection of the present invention.

Claims

1. A smart management system for billiards matches based on AI visual analysis, characterized in that, include: High-definition camera: Installed above the billiard table to capture video streams of the game; AI edge computing box: Built-in GPU / NPU, running deep learning models for real-time analysis of the competition video stream captured by the high-definition camera; Billiards recognition model: a ball detection and trajectory tracking model trained based on YOLO or DeepLab algorithms; Rules engine module: Performs logical judgments based on the rules of billiards competition; Mobile application / large screen system: used to display real-time scores, highlights, and match data; Storage server: Used to store video clips, match logs, and highlight markers.

2. The intelligent management system for billiards matches based on AI visual analysis according to claim 1, characterized in that, The AI ​​edge computing box receives the RTSP / H.264 video stream from the high-definition camera and processes it through the following steps: Video preprocessing: Denoising, distortion correction, and illumination compensation are performed on the video stream images captured by the high-definition camera; Sphere detection: Using a CNN model to identify the positions of 15 target spheres and the white sphere; Trajectory tracking: Tracking the trajectory of a sphere using SORT or DeepSORT algorithms; Pocket entry recognition: Combines pocket opening area detection with ball disappearance logic to determine a goal event; Foul detection: Identifies actions such as the cue ball being pocketed, not touching one's own ball, and jump ball fouls; Spectacular Shot Judgment: Scoring is based on three dimensions: difficulty of scoring, number of consecutive hits, and positioning accuracy, and the shot is marked as a spectacular moment.

3. The intelligent management system for billiards matches based on AI visual analysis according to claim 1, characterized in that, The workflow of the AI-based visual analysis-based intelligent management system for billiards matches is as follows: Intelligent scoring and refereeing: After the AI ​​edge computing box identifies a goal, it automatically updates the current player's score. After a foul is determined, the ball possession is automatically switched, and the type of foul is indicated. Real-time score synchronization to mobile applications and large-screen systems enables intelligent referee functionality; Highlight Recognition and Replay: When the system detects events such as clearing the table, potting a long ball, or solving a snooker, it automatically marks the timestamp. Users can quickly view the highlights recommended by AI by clicking the "Highlights" button in the mobile app; It supports adding special effects to enhance the viewing experience, such as slow motion, spotlights, and sound effects; Match statistics: Automatically generate match reports, including: number of goals, success rate, number of fouls, average shot time, etc. Supports historical match review and data comparison analysis.

Images