Edge Detection

This lecture summary deals with the low level image processing task of edge detection. Edges are discontinuities, significant local changes, in image intensities which arise from three sources: (1)projection of 3D contours (2)texture present on the 3D surfaces (3)shadows cast by the imaged objects. The summary includes the classical derivative based operators due to Canny [2], and Marr [7] and one non derivative based operator SUSAN [9]. We will not deal with the subject of algorithm evaluation. People which would like to read about this subject are referred to [1, 5, 8] evaluation studies of edge detection algorithms according to different criteria. The summary is divided into three sections: (1) Derivative based operators. (2) The SUSAN edge detector. (3) Post processing.