Corner Detection Based on Distance, Angle and Template
Yang Haifeng · Laser & Infrared · 2005
A corner detection algorithm based on extrema of a distance sequence and a angel sequence is proposed, which is distinct from other approaches based on curvature and intensity, and also overcomes some defects of them. Firstly, extrema detection of a distance sequence and anangel sequence are applied to find corners of apparently turning parts, and then corners of gently turning parts are obtained by distance template marching. At last, the final result is gained after deleting false corners. Experiments show that this algorithm restrains influences of the quantization error and noise by controlling sizes of the distance template and supporting region of extrema in sequences,and it has high precision and stability of the corner detection of 2D image and location.It is a practical, effective corner detection algorithm.