Effect of Modified Gauss Mixture Model on the Processing Efficiency and Accuracy of Moving Objects in Video of Sports Matches

Aiming at the drawbacks of current sports video moving target detection, this paper proposes an improved Gauss mixture model for sports video moving target detection and tracking. By analyzing the disadvantage of the Gauss mixture model, the original process of "background reconstruction model updating background updating target detection" is retained. The color image is transformed into gray image, and the background area with small difference in pixel similarity is expanded dynamically. The target tracking process based on gray histogram is added to improve the processing efficiency and accuracy of Gauss mixture model in sports video moving target. The experimental results show that the proposed method has strong ability to deal with sports video interference, large detection range and good detection and tracking effect.