Table Tennis Falling Point Recognition And Scoring System Based On Object Detection And Tracking

Technology is changing lives.With the increasing use of image processing technology in sports games,in ball games,the landing point recognition system for the identification and positioning of fast-moving spheres plays an importan t role in games and daily training.It has become the current One of the resea rch hotspots of artificial intelligence in the field of entertainment.The table ten nis drop point recognition and scoring system based on target detection and tra cking designed in this paper focuses on the application of target detection and tracking technology in the intelligent table tennis training scene.By collecting athletes 'playing videos,the table tennis drop point is evaluated.Accurately ide ntify and analyze the area of landing,and record the training effect of athletes.The table tennis drop point recognition and scoring system based on target detection and tracking designed in this paper adopts a client / server(C / S)distributed design model,in which the client is responsible for human-computer interaction,including collecting user voice data,Perform semantic analysis,sen d training control instructions,transmit playing video data,and be responsible f or visualizing recognition of landing points and playing scores.The server is r esponsible for receiving voice commands and parsing the commands,receiving the video stream transmitted from the client,performing table tennis detection and tracking on the training video,performing spot recognition and area analys is,and finally returning the results to the client.In order to identify the falling point of table tennis,this paper improves t he traditional Vibe target detection algorithm,including: increasing the neighbor hood sampling range of the algorithm and reducing false detection;adding an adaptive background change threshold selection to improve detection stability;d etection area Continuous area filtering and pixel flicker point detection to remo ve noise interference.In the tracking stage,in view of the shortcomings of the KCF tracking algorithm in complex table tennis training scenarios,this paper has designed a reasonable method of combining detection and tracking to judgethe loss of tracking in the next frame of the video.If the target is lost,Then the tracking was abandoned and the improved detection algorithm was enabled to meet the basic needs of measuring table tennis coordinates in real time.In the trajectory reconstruction stage,this paper proposes a reasonable division me thod of serving and receiving rounds,using the change of the slope of the pin g-pong trajectory to realize the judgment of the ping-pong "touch table",and d esigning an accurate trajectory equation fitting scheme.This paper proposes a s ystematic scoring scheme for the playing area.The system algorithm has carrie d out a number of experiments in the table tennis intelligent training system of the project team,accurately and effectively performing table tennis drop point recognition and area scoring,and basically achieved the expected design require ments.