一种改进的角点探测方法  

FAST CORNER DETECTION ALGORITHM

作　　者：朱如军[1] 孙先仿[1] 

机构地区：[1]北京航空航天大学自动化和电器工程学院,北京100083

出　　处：《计算机应用与软件》2005年第1期105-106,129,共3页Computer Applications and Software

摘　　要：提出了一种改进的角点提取算法。该算法是基于角点的性质 ,即在角点的某个邻域内 ,图像灰度变化在任意一条通过该点的直线上都很大。本文利用此性质对象素点直线方向上灰度变化值求取极值 ,建立角点候选点。再通过SUSAN特征检测原则进一步剔除虚假角点。测试表明改进后的算法具有运算量小 ,定位精度高 ,抗噪能力强的特点 ,易于实时实现。This paper proposes an improved corner detection algorithm,based on the property of corners that the change of image intensity must be high in all directions.This algorithm gets potential corners by calculating the extremum of the intensity changes of the image with this property,then uses SUSAN principle for feature detection to get rid of false positives.After these two steps,it uses non minimum suppression to reserve true corners.The experiments show that this method has low computational complexity,high accuracy,and it is not sensitive to noise.It was found that this algorithm was suitable for real time jobs.

关 键 词：角点 SUSAN 特征检测 提取算法 象素 抗噪能力 灰度 直线 运算量 性质 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] G633[自动化与计算机技术—计算机科学与技术]