一种改进的Shi-Tomasi角点检测方法  

AN IMPROVED SHI-TOMASI CORNER DETECTION METHOD

作　　者：彭晓星 罗霄 韩宝玲[3] Peng Xiaoxing;Luo Xiao;Han Baoling(School of Optics and Photonics,Beijing Institute of Technology,Beijing 100081,China;School of Computer Science&Technology,Beijing Institute of Technology,Beijing 100081,China;School of Mechanical Engineering,Beijing Institute of Technology,Beijing 100081,China)

机构地区：[1]北京理工大学光电学院,北京100081 [2]北京理工大学计算机学院,北京100081 [3]北京理工大学机械与车辆学院,北京100081

出　　处：《计算机应用与软件》2022年第12期241-245,共5页Computer Applications and Software

基　　金：国家重点研发计划项目(2016YFC0803000,2016YFC0803005)。

摘　　要：针对Shi-Tomasi角点在单一尺度下角点提取精度低、易提取伪角点、抗噪性差的问题,提出一种改进的Shi-Tomasi角点检测方法,通过构建图像金字塔实现在不同尺度下Shi-Tomasi角点的提取,形成候选点集,对候选角点采用双模板组合的方式去除伪角点,实现精提取。通过实验证明该算法可以在多尺度下提取角点,且增强了图像的抗噪性以及提升了提取角点的精确性。In order to solve the problems of low accuracy of corner extraction of Shi-Tomasi corners at a single scale, easy extraction of pseudo-corner points and poor noise resistance, this paper proposes an improved method of Shi-Tomasi corner detection. By constructing an image pyramid, Shi-Tomasi corner points were extracted at different scales to form a candidate point set. Pseudo-corner points were removed by using a dual template combination method to achieve fine extraction. The experiments show that this algorithm can extract corners in multiple scales, and enhance the image noise resistance and improve the accuracy of corner extraction.

关 键 词：Shi-Tomasi角点 多尺度角点 高斯图像金字塔 双模板 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]