Robust Edge Detection Method with Subpixel Accuracy in Presence of Noise

Edge detection is one of the most important issues in image processing and machine vision. Edge detection in image processing is a low order process, so that the performance of the higher order processes such as object identification, segmentation and coding of images is directly related to the efficiency of this process. The estimation of edge parameters with using gradient vector calculation is usually not accurate. Keeping the structure of edge is one of the most important problems in edge detection, especially in detecting noisy images. For practical applications that accurate edges are needed, subpixel edge detection is done. In this paper a new edge detection method based on edge figure and obtained model from neighboring pixels effect and spatial relation of image pixels is introduced. Then an iterative restoration process based on presented edge detector is suggested. The purpose of this method is to increase the accuracy in recognition of subpixel position, curvature, orientation and change in intensity in noisy images.