Corner detection based on distance transform and morphology

A simple and effective corner detection algorithm based on morphological and distance transform is proposed. In essence, the algorithm is different from the corner detection algorithm which is based on chain-code, oriented edge or morphology. It overcomes some bugs of former algorithms from basic principle. At last, some corner detection experiments of binary and intensity images are given. Experiments indicate that the algorithm overcomes some shortcomings of other algorithms and preserves their advantages. It is a simple, practical, effective corner detection algorithm.