适于光学遥感图像的角点特征检测算法

Algorithm for corner feature detection of optical remote sensing image

角点特征检测和精确定位是图像配准与匹配的基础。探讨了Moravec算子和Harris算子的基本原理,并进行了图像的角点特征检测和实验分析,结果显示Harris算子效果要好。但是,Harris算子对真实遥感图像角点特征的检测尚不能取得理想的结果,因此,对Harris算子进行了改进。实验表明,通过改进后的Harris算子对遥感图像上的角点特征的检测具有较好的检测效果。

Corner feature detection and accurate location are the foundation of image registration and matching.This paper discusses the basic principle of Moravec operator and Harris operator,and detects the corner feature of the image and analyzes the test results,the results show that Harris operator gets better results.However,Harris operator can not achieve the desired results on detecting the corner feature of the real remote sensing image,therefore,this paper improves Harris operator. The experiments show that improved Harris operator can obtain good test results on detecting the corner feature of the remote sensing image.