A Directional Edge Detector Operator to Detect All Type of Edges of Image

Abstract

This work proposes a new directional edge detector operator to detect all type of edges of the image. Initially this thesis provides an introduction to digital image processing followed by a review of various edge detection methods. In the research part authors initially did few experiments on the famous Lena image then by these experiments authors come to two conclusions. These conclusions are basic pillars of the proposed work. (a) Image contains smoothly varying area separated by edge information. (b) Edge pixels take intensities comparatively smaller or larger than the surrounding pixels. Based on this conclusion authors have design the simplest method that works on image patches and find edge pixels. . This operator works on information aligned in four different directions. With the help of a mean filter or a given threshold the proposed method is able to detect strong edges of the images.