A Novel Method for Corner Detection Based on Conditional Skeleton
一种基于条件骨架的拐点检测方法

A source image is regarded as a polygon, and the extended line of the morphologic conditional skeleton must cross the vertex of a polygon. The convex corners can be obtained through detecting the zero radius of the maximum plate on morphologic skeleton. All the results of corner detection are given by logical hetero or operation between two corner sets of source image and its supplemental image. In addition to the pseudo lattice, corners which are generated by discrete sampling can also be eliminated.This method has very high value in exactly locating the corner,can reduce the computational cost,only requires simple hardware equipment,and is convenient for parallel processing.Because of using large plate to detect corner,this method is good for noise filtering and can be popularized to gray images by using two kinds of large plate.