Optimal Design of Video Player Tracking Algorithm in Basketball Game Based on Support Vector Machine

With the rapid growth of the number of videos, how to efficiently retrieve the required video information has become an urgent problem, and video annotation is the basis of video indexing and video search. Target tracking technology is of great significance in basketball game video, which is the basis of high-level semantic tasks such as video summary generation, player movement analysis, game strategy formulation and basketball event detection. Intelligent video monitoring system mainly refers to the automatic detection, segmentation, tracking, classification and higher level behavior analysis of objects in the monitoring scene without human intervention, so as to achieve the goal of “real-time monitoring and timely response.” In this paper, a real-time player segmentation and tracking algorithm is implemented to solve the problem of player team membership and tracking in basketball video. Compared with support vector machine algorithm, the two algorithms proposed in this paper have better robustness and accuracy.