一种改进的快速SUSAN角点检测算法  

Improved Algorithm Based on SUSAN Corner Detection

作　　者：林鹏岳[1] 李玲玲[2] 李翠华[1] 

机构地区：[1]厦门大学计算机科学系,福建厦门361005 [2]郑州航空工业管理学院,河南郑州450015

出　　处：《计算机与现代化》2010年第2期66-68,72,共4页Computer and Modernization

基　　金：国家重点基础研究发展计划(973计划)资助项目(2007CB311005);福建省自然科学基金计划资助项目(A0710020);河南省基础与前沿技术研究计划资助项目(092300410043)

摘　　要：着重分析SUSAN角点检测速度慢的缺点,提出改进的快速SUSAN角点检测算法。该算法首先比较待测点与其周围半径为R的圆周上像素的灰度,统计连续的具有非相似灰度值的像素个数,根据判定准则得到初始角点,再使用非极值抑制函数选择最优角点。实验结果表明:该改进算法有效地减少角点检测过程的计算量,使得角点检测的效率大大提高。The reason for the low speed of SUSAN algorithm in comer detection is analyzed, and an improved fast comer detection algorithm based on SUSAN algorithm is proposed. First of all, the algorithm compares the gray value of current point with the gray values of the points with a radius R around it, and counts the number of pixels with a non-similar gray value. The initial comer points are marked according to the criterion presented in the article. Then, a non-maximal suppression function is used to choose the best comer. Experimental results show that the improved algorithm greatly reduces the computation of comer detection process, and makes efficiency improved.

关 键 词：SUSAN算法 角点检测 快速SUSAN算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]