The application of Hough Transform for corner detection

Corners are very attractive features of many applications in human perception and computer vision. The problem of detecting corners is simplified into detecting simple lines in a local coordinate system. As simple lines can be expressed using only one parameter, Hough Transform can detect them quickly. Other edge detectors have some problems when marking edge points around corners. Based on both gradient magnitude threshold and grey level analysis, the edge points can be detected. Using the Sobel and Kirsch Operator, an image can be extracted and provide the edge points for Hough Transform. Mean while, the Zhang Suen Thinning Method can be used to reduce the edge points into minima points to speed up the algorithm. To illustrate the problem, the interface program is developed by using Microsoft Visual C++ 6.0