A novel algorithm of detecting corner points for the contours of lines and arcs

Inhalt:
A novel algorithm is proposed to extract corner points for the contours of lines and circle curves. The corner points are important to segment contours into line segments and arcs for further processing. The algorithm first fit LC curve primitives and then calculate the nearest distances between the set observation points and the fitting primitives; the extreme points of the nearest distances can be easily detected as corner points. Experiment results show the proposed algorithm is efficient and effective.