Edge detection and extraction from noisy image

Includes bibliographical references (page 45)Edge in noisy image can be detected by thresholding\ud the outputs of differentiator, after image enhancements.\ud The object of this project is to develop algorithms that\ud can be used for detecting and extracting edges from noisy\ud image by a personal computer. The raw image is first\ud subjected through a 5-step image processing procedure:\ud (1) smoothing; (2) compass gradient filtering; (3) thresholding;\ud (4) thinning; and (5) edge extraction. Finally,\ud edge directional parameter is used as a feature for object\ud recognition