基于局部区域像素相似度的种子图像边缘检测

he image edge is the most basic feature of the image, and it is the result of gray discontinuity. The edges of the image significantly reduces the amount of data and keeps important structural properties of the image. Edge detection is the first step in image analysis and understanding. The edge information of the image is of vital importance in a variety of image transformation and processing. This paper introduces an edge detection algorithm based on local grayscale comparison．How to select a template and threshold in this algotithm are discussed. This edge detector was realized in Matlab. The experimental results show that this edge detector is superior to traditional edge detection methods on images with noise. The algorithm is suitable for edge detection in noisy images or low-contrast images.