A fast detection algorithm of Harris apparent corners based on the local features

A fast apparent corner detection algorithm based on the local features is proposed in this paper. Firstly, the pixels on each side of the grid region are stretched out, and the edge points of the object on each side are detected according to the differences of the gray values of the pixels. The number of the edge points is employed to decide whether the grid contains a corner or not. Then the candidate grids that do not contain a corner are removed via checking the angle between the edges of the object. The method of average position is used to exclude gathered corners. Then the remained corners are taken as candidate corners. Finally, the Harris algorithm is applied to the small area around each candidate corner to get the apparent corners. Experimental results verify the effectiveness of the proposed algorithm.