双阈值Harris角点检测算法  

Double threshold value Harris corner detection algorithm

作　　者：吴学礼[1,2] 丁雪[1] 孟凡华[1,2] 赵萌[1] 

机构地区：[1]河北科技大学电气工程学院,河北石家庄050018 [2]燕山大学电气工程学院,河北秦皇岛066000

出　　处：《河北科技大学学报》2013年第4期330-333,共4页Journal of Hebei University of Science and Technology

基　　金：国家自然科学基金(60974018;61273188);河北省自然科学基金(F2012208075)

摘　　要：角点检测是图像处理中的一种重要的特征提取方法,为了改进角点检测算子的检测性能,提高角点检测的精度,降低检测出伪角点的概率,针对Harris角点算法的阈值受人为给定的限制的缺陷,提出了一种基于双阈值的Harris角点检测,利用小阈值T1来检测角点,利用大阈值T2来限制伪角点的出现,并通过大量实验总结出大小两阈值的关系,获得良好的检测性能。Corner detection is a kind of important characteristic extraction method in image processing. In order to improve the detection performance of corner detection operator and hence improve the accuracy and reduce the probability of detecting false corner detection, a method based on dual-threshold Harris corner detection algorithm is presented to overcome the defects of Harris corner threshold values affected by artificial setting limits. The small thresholds are used to detect the corner points, and the large thresholds are used to limit the emergence of false corner points. The relationship between the two threshold val- ues is summarized through a large number of experiments to obtain good detection performance.

关 键 词：HARRIS 角点检测 双阈值 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]