摘要

The scientific analysis of athletes' training images in competitive sports training can effectively help improve the training effect. This paper carries out the analysis aimed at the image processing technology with the collected video images of sports; the images are filtered to obtain the preprocessed images, which are grayed up for the qualitative analysis. The processing of sports images requires real-time tracking of the athletes and the sports equipment. To solve the shortcomings of traditional algorithms, this paper proposes a nonlinear detection algorithm based on the research of current experts and scholars. The algorithm mainly determines the tracking object by measuring its boundary through the maximum and minimum value in the same pixel neighborhood. Finally, new training strategies can be developed on the basis of the data to enhance the skills of athletes.