An Adaptive Corner Detection Algorithm Based on Edge Features

Abstract

Harris algorithm is classical and hight speed corner detection methond, but the detect result is dependent on the thresholds and the coefficients of the corner response function, and it is easy to lose and error detect corners, so it has lower accuracy of corner detection. In summary, an adaptive corner detectijon algorithm base on edge feature was proposed in this paper, the purpose of improve detection algorithm is to promote corner detection accuracy and ensure algorithm performance. Firstly the Canny operator is used to extract the edge information of image, and calculated the characteristic value of edge region is calculated by optimizing the corner resopnse function; then, the adaptive threshold is obtained by the OSTU algorithm to extract candidate corner region; Finally, according to block detection and scale-invariant ideas, we use the non-maximum suppression algorithm is improved to remove false corners. Our proposed algorithm not only overcomes somewhat non-adaptive threshold and easily loss corner and error detection problem, but also improves speed detecion. The experimental results show that our algorithm can improve the accuracy of corner detectijon, and can have ideal detection effect in different envionments.