基于SIFT特征和广义紧互对原型对距离的遥感图像配准方....pdf

文档介绍：基于 SIFT 特征和广义紧互对原型对距离的遥感图像配准方法
霍春雷〔1〕,周志鑫〔2,1〕,刘青山〔1〕,卢汉清〔1〕
1 (中国科学院自动化研究所模式识别国家重点实验室,北京,100080)
2 (北京市遥感信息研究所,北京,100854)

摘要:本文主要讨论 SIFT(Scale Invariant Feature Transform)及其在遥感图像配准中的应用。首先介绍了基于特征点的遥感图
像配准的一般框架;针对基于特征点的遥感图像配准中的两个基本问题:鲁棒的特征点提取和特征点匹配,本文提出了基于
SIFT 特征点和广义紧互对原型对距离的遥感图像配准新方法,并通过“广义紧互对原型对”的概念,为不同的特征点匹配
方法建立了联系。与已有的相关工作相比,该方法可以得到更多的匹配点对和正确的匹配点对。数值试验证明了该方法的有
效性和鲁棒性。
关键词:遥感图像配准;SIFT;广义紧互对原型对;特征点匹配
中图分类号: 文献标识码:A

Remote Sensing Image Registration Based on SIFT and the Distance Between
Generalized Tight Pair-wise Prototypes
HUO Chunlei〔1〕, ZHOU Zhixin〔2,1〕, LIU Qingshan〔1〕,LU Hanqing〔1〕
1 (National Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing, 100080)
2 (Beijing Institute of Remote Sensing, Beijing, 100854)

Abstract: SIFT(Scale Invariant Feature Transform) and its application on remote sensing image registration is
discussed in this paper. First, the keypoint-based framework of remote sensing image registration is described;
After analyzing robust feature extractor and keypoints matching which are two key problems in the image
registration, a new approach based on SIFT and the distance between generalized tight pair-wise prototypes for
remote sensing image registration is proposed; The relation between different matching methods based on SIFT is
established by the c