Technique Of Automatic Judgment Of Offside In Soccer Games Based On Video Analysis

In our soccer games, it can not ensure fair competition only by manual penalty, but the rapid development of multimedia technology makes it possible to judge the game automatically by machine. According to the frequent phenomenon of offside misjudgment in the soccer game, this paper takes the soccer video as an entry point and proposes a solution of automatic offside judgment on the basis of profound acknowledgement on the offside rules.The coordinates of the ball and players can be precisely obtained and reconstructed through computer vision and video image processing technology. On this basis we design correlation algorithm of offside reasoning and realize it by programming. This paper establishes a framework for offside automatic judgment system, and then divides the chapter to describe the key technology that the judgment needs. Firstly, we design the algorithm of detecting the moving objects in soccer video. In playfield detection, the HSV color model transfer is utilized. And in the player detection, the background modeling is adopted according to the playfield, and then the non-court areas are removed by region analysis. In this way the problem of interference and players’ interceptions are solved by designing the players’ parameter templates. In the ball detection, the Gauss template is designed and the ball estimation function is introduced according to the size and the shape features of ball.At the same time, this paper also discusses the key technology about the scene analysis and reconstruction. Camera calibration theory is utilized to reconstruct the playfield plane coordinates. The camera perspective projection is adopted to solve the projection matrix after substituting the playfield feature points. Then the image coordinates of ball and players are converted into the world coordinates. On this basis the algorithm of digital offside judgment is designed and implemented by simulation. The experimental results show that the algorithm proposed in this paper is practicable and effective, thus it can be adopted as electronic referee to assist the judgment.