Tennis Serve Trajectory Capture Algorithm Based on Wavelet Multiscale Decomposition

Human motion analysis is one of the important research directions in computer vision technology. Aiming at problems poor effect, accuracy, and efficiency current tennis serve trajectory capture process, a algorithm based on wavelet multiscale decomposition proposed. This article analyzes related concepts principles transform multiresolution analysis, uses three-dimensional acquisition method to collect image, mean filter smooth denoise collected image. The used process image after denoising, find local maximum point through modulus value phase angle value, obtain initial Using layer threshold method, adaptive set, line under each scale obtained serve. It can be observed from experimental simulation results that effect proposed good, it effectively improve accuracy capture.