Gray-level Corner Detection by Virtual Electric Field Model

Corner detection and feature extraction are widely used in image processing and computer vision applications. In this paper, we propose a new gray-level corner detection based on virtual electric field model having different configurations, which demonstrate the best detection. The basic idea is to model the image as a grid of virtual electric charges of a plane surface in electrostatic balance. The algorithm is achieved in two main steps. The first one uses an adaptive method for computing directional gradients which permits to build an edge detection operator of 3times3 pixels. The second one introduces this operator for determining image derivatives in order to detect the corners. The smoothed squared images derivatives are given by using a Gaussian function. Then, the response function proposed by Harris and Stephens is computed for each pixel, using auto-correlation matrix. The local maxima search is performed as a final step of the method. Simulation results on synthetic and real images show the performances of the proposed method compared to that presented by an improved version of Harris detector.