Aiming at the problem that the traditional Canny edge detection algorithm is sensitive to salt-pepper impulsive noise and the small change of edge gradient magnitude

so that it cannot extract the target contour edge information effectively

an image edge detection algorithm based on an improved Canny operator is proposed in this paper. The new algorithm starts with multi-level threshold asymmetric clipping median filtering based on neighborhood similarity to replace Gaussian filtering

adaptively selects filter windows for noise reduction based on the neighborhood similarity principle and decision-based multi-rule; then extracts gradient vectors by convolutional templates in eight directions to improve edge localization accuracy; and finally uses the maximum inter-class variance (Otsu) algorithm to calculate the threshold value to improve the edge detection adaptivity. The experimental results show that the denoised image quality is improved by 10% to 20% and the edge evaluation index is improved by 15% to 20% compared with the traditional edge detection operator. The algorithm can effectively remove the pepper noise and has a better edge extraction effect.