A review of classical and fractional derivative order methods for edge detection in image processing

Abstract

The main goal of this work is compare between the classical edge detectors and an approach of edge detection based on fractional differentiation. Classical edge detectors based on first order derivative: Prewitt, Sobel and Canny and one edge detector based on second order derivative: Laplacian is discussed. Also the 2nd order fractional differentiation approach of edge detection is presented. To check the effect of image smoothing on edge detection results additionally, for results comparison between different methods, a smoothing filter is applied for Prewitt, Sobel, Laplacian. The comparison is conducted from time consuming point of view and quality of edge detection level.