Research On Analysis Method Of Player Motion Data In Soccer Match Video

Through the analysis of soccer game videos,the players’ sports data are obtained,and statistical analysis and visual display are carried out,which is of great significance for increasing the audience’s viewing experience and assisting coaches to evaluate the players’ performance.Compared positioning devices,using computer vision technology to obtain player motion data can greatly reduce the cost,and it is also very easy to expand to other motion scenes.Based on the player’s sports data,it can assist referees in making decisions,assisting lens event detection,etc.In order to solve the problem of less research content in player movement data statistics,a method is designed to obtain player movement data and carry out statistical analysis by using court area detection,court reconstruction and player tracking.The obtained player movement data includes position,speed,trajectory,heat diagram,distance between players,etc.Aiming at the interference problem of non-stadium area in soccer field analysis and processing,a stadium segmentation algorithm based on dominant color is designed,which reduces the influence of non-stadium area in straight line detection.Aiming at the calculation of homography matrix between game video frame and standard two-dimensional stadium,The soccer field area is divided into three parts: left,middle and right.For the left and right parts of the stadium,the homography matrix is obtained by detecting straight lines and grouping them,and the matching relationship between straight lines and intersection points is used.For the midfield area of the stadium,the homography matrix is calculated by identifying the matching relationship between the middle circle and the key points of the middle circle.Aiming at the statistical problem of player movement data,an algorithm is designed to calculate and display player movement data by using homography matrix and player tracking coordinates.Experimental results in soccer far-shot video sequences show that the homography matrix is used to map the stadium and calculate the intersection-parallel ratio for evaluation,and the intersection-parallel ratio can reach 87%.Based on this mapping transformation,the player’s motion data can be better calculated and displayed.Because all straight lines on the soccer field cannot be detected,At the same time,pixel-by-pixel recognition of the middle circle leads to a long time-consuming algorithm,and the calculated and statistical player data is not complete enough.Future work will focus on improving the detection of the court line,optimizing the middle circle recognition algorithm,further improving the player data and trying other methods to calculate the court mapping relationship.