Moravec和Harris角点检测方法比较研究  

Research on Moravec and Harris Corner Algorithm

作　　者：卢瑜[1] 郝兴文[2] 王永俊[1] 

机构地区：[1]华南师范大学数学科学学院,广东广州510631 [2]潍坊学院数学与信息科学学院,山东潍坊261061

出　　处：《计算机技术与发展》2011年第6期95-97,100,共4页Computer Technology and Development

基　　金：国家自然科学基金(61075033;61005033)

摘　　要：角点是图像中的重要特征,在图像配准与匹配、运动分析、目标识别、目标跟踪等领域中均得到了广泛的应用。Moravec和Harris算法是计算机视觉领域中应用比较广泛的角点检测方法。介绍了Moravec和Harris及其改进等四种常用的角点检测算法的实现原理及其特点,并对这四种方法进行了分析、比较,给出了它们的效率性能评价。然后通过实验研究了它们在角点检测的速度、正确率和抗噪性等方面的差异,探讨了各种算法的优势和缺陷。最后,文章指出了角点检测技术的研究与发展方向。Corner is an important feature,which has been used in image registration and matching,motion estimation,object recognition,target tracking.Moravec algorithm and Harris algorithm are two of the most widely used corner detecting methods in computer vision community.In this work,Moravec algorithm and three Harris algorithms are studied,categorized and compared.The performance evaluation of corner detection was investigated.Compared them with each other and evaluated them on efficiency,precision,accuracy.The advantages and disadvantages of them were also investigated.Finally,point out the future research and development direction of corner detection.

关 键 词：角点检测 HARRIS算法 角点响应函数 阈值 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]