Edge Detection of Strong Noise Image Based on WMF and Improved Canny Algorithm

Abstract

Traditional edge detection algorithm is not effective in detecting images with strong noise. When the image to be detected is heavily polluted by salt and pepper noise, traditional edge detection algorithms have a fairly high false detection rate and missed detection rate. To solve this problem, this study proposes an improved algorithm based on WMF (weighted median filter) and an improved Canny algorithm. This method determines the size of the filter window adaptively according to the number of noise points, assigns the weight level according to the degree of membership of the pixels in the window, and then uses the WMF algorithm to de-noise the image. A dual global threshold selection algorithm is used to adaptively determine the high and low thresholds to reduce the detection problems caused by artificial thresholds. The simulation results show that the edge extracted by the proposed algorithm is better than that extracted by the classical algorithm under the strong noise of salt and pepper. Although the detection time is slightly increased and the operating efficiency needs to be improved, the detected effective edge is significantly more than that of the classical canny algorithm.