基于改进边缘检测算子的图像特征点提取算法  

Algorithm for Image Interestring Point Detection Based on Improved SUSAN Operator

作　　者：郭峰[1] 王道平[1] 韩溟[1] 

机构地区：[1]第二炮兵工程学院电工电子技术教研室,陕西西安710025

出　　处：《电子科技》2011年第9期9-11,共3页Electronic Science and Technology

摘　　要：介绍了边缘检测算子的原理,对算法中灰度差阈值的选择方式进行了改进。在图像中每个像素的SUSAN模板中首先计算阀值t的初值,再利用迭代法获得终值,使其在各种不同的对比度下仍能正确检测兴趣点,最后运用该算法进行了图像边缘检测测试,并与其他检测算子进行比较。实验结果表明,该算法具有方法简单、抗噪能力强、计算量小等优点。SUSAN edge detecting operator is an edge detection algorithm based on comparison of the image gray.This article introduces the principle of the operator,and improves the gray Threshold.The initial value of threshold is calculated in every single SUSAN template.Final value of threshold is obtained using iterative segment,which makes it possible to pick up the right interesting point in all kinds of gray contrast.The results show that the algorithm has the advantages of simplicity,strong anti-noise,and small amount of calculation.

关 键 词：SUSAN算法 兴趣点检测 迭代 自适应 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]