Image Edge Detection by Mean Difference Thresholding

Abstract

Edge detection is very essential step for numerous image processing tasks. In this paper, the different techniques of edge detection, i.e., Roberts, Prewitt, Sobel, Isotropic, Laplacian of Gaussian (LoG), Difference of Gaussian (DoG) and Canny are discussed and also the diagrammatic difference between the some of these techniques are reviewed. This paper describes a simple edge detection technique based on thresholding, which is used to make fast process of edge detection and represents results in more efficient way. The validation of proposed algorithm is done by MSE and PSNR parameters. There are many edge detection techniques available, but any single technique is not suitable for all the applications due to the variation in quality of an image and the variability in shape of an image. So, the selection of the technique is dependent on the requirement of the applications.