Abstract: In order to improve the accuracy and stability of circular image feature point detection,an algorithm is put forward based on the gray-scale symmetry of circular image center.Through this algorithm,the edge of circular image is firstly detected by Sobel operator;then gray-scale qualitative method is used to determine the center of the circular image;and finally gray-scale symmetry factor is introduced to obtain the sub-pixel coordinates of the feature points of circular image.Projection simulation experiment and measuring experiment are adopted to evaluate the precision of this algorithm,and the experiment results show that the precision of this method is around 0.2 pixels.

Key words: feature points, sub-pixel, corner detection, camera calibration