Target Detection and Tracking Technology of Aerobics Sports Training Based on Hybrid Algorithm

Abstract

In order to make the aerobics training method more scientific and reasonable, the training features and skills in the video can be extracted through the moving target detection (MTD) and tracking technology (TT). However, in aerobics sports videos, the speed of the moving target is often faster than that of the static target, and the background is more complex, and a single tracking algorithm (TA) cannot play a good role. Therefore, this paper proposes a hybrid tracking algorithm based on the background difference method (BDM) and Kalman filter. The hybrid algorithm models the background of the video image, improves the detection accuracy, and has a higher tracking accuracy than the single algorithm.