Signal abrupt point detection based on wavelet transform

This paper investigates a method for detecting signal abrupt points based on wavelet transform. An algorithm is proposed to locate abrupt points using the extrema of wavelet decomposition coefficients and eliminate short branches. By comparing the performance of different wavelet functions, the Haar wavelet with a scale of 4 is selected for detection. Experimental results show that the proposed method accurately detects and removes skeleton fold points, demonstrating high accuracy and practicality.