Image edge detection and segmentation based on the Hilbert transform

Abstract

An efficient edge detection model based on a two-stage procedure is presented. The first stage consists of a linear transformation of an artificially created signal that has the property of detecting step edges irrespective of the distribution of the intervals between two consecutive edges. It is demonstrated that the Hilbert transform possesses this property and also outperforms the derivative operation in the detection of step edges in the presence of noise. The second stage consists of an appropriate mapping of the filtered results into edge detection primitives. Practical confirmation of the results is given by means of examples.<>