A Novel Edge Detection Algorithm Based on Distance

Abstract

Abstract Under the framework of Canny algorithm, a distance-based edge detection algorithm is proposed to improve the gradient magnitude of Canny. In our algorithm, the gradient magnitude can be acquired by taking the distance from the center of the mask as the weight factor. The operator not only can calculate the image gradient better, but also has good separability in the horizontal and vertical directions, and its gradient amplitude has a certain degree of rotation invariance. Finally, this operator is compared with various gradient operators for the Lena image simulation and the actual waterfront edge detection experiment. It shows that the operator has a better edge detection effect from the experimental results.