基于方向性SUSAN算子的图像角点特征提取  

Corner Features Extraction of Image Based on Orientation SUSAN Operator

作　　者：贺菁[1] 李庆华[1] 王新赛[2] 

机构地区：[1]华中科技大学计算机科学与技术学院,湖北武汉430074 [2]防空兵指挥学院红外与射频技术研究中心,河南郑州450052

出　　处：《小型微型计算机系统》2008年第3期508-510,共3页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(60273075)资助

摘　　要：在不同传感器图像的匹配中角点是重要匹配特征,为满足实际一些具有方向性的成像传感器图像匹配的需要,进行稳定可靠、具有方向性角点特征的检测是必需的.本文采用方向性的SUSAN角点检测算法,将SUSAN算子计算区域限定在具有一定指向的扇区,使之具有方向性检测角点特征.实验证明,该算法简单、实用和有效.The corner feature is an important character for image matching of the different sensors, the corner is defined as the of very high curvature on image edge commonly. In order to meet practice need of image matching for imaging sensor ntation, it is necessary to detect corner feature that is stabilization, credibility and orientation. In this paper the algo-orientation SUSAN corner detection is adopted, the method is the calculating region of SUSAN operation limited an appointed sector, so that corner feature is detected with orientation. Experiment indicates that this method is simple, practical and effective.

关 键 词：方向性SUSAN算子 角点检测 特征提取 图像匹配 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]