基于多源遥感影像融合的影像匹配技术  

Image matching based on multi-source remote-sensing image fusion

作　　者：刘晓龙[1] 李英成[1] 

机构地区：[1]中国测绘科学研究院,北京100039

出　　处：《测绘科学》2007年第3期59-61,共3页Science of Surveying and Mapping

基　　金：国土资源部科研项目-土地利用动态遥感监测作业系统开发项目

摘　　要：多源遥感影像数据配准,常用的方法是多项式纠正法。此方法简单,但不能有效实现图像之间的相互配准。本文介绍了一种基于影像匹配的图像对图像局部纠正技术,包括影像直方图匹配、特征点提取、影像匹配处理技术,以及图像对图像局部纠正等技术,用于多源数据的配准。实验证明这种流程适合不同时相、不同传感器遥感数据(TM数据、SPOT数据与航空影像数据)的精确配准,配准误差可达到子像元级别。Polynomial rectify is usually used for relative correction technology of multi-source remote sensing data. It is simple but inefficient. Based on image matching, a local correction technology of image to image is introduced in this paper. This local correction technology includes histogram matching, characteristic point extraction, image matching and processing, etc. So it is used for multi-source data registration. From many times experiments, it is showing that this technology is useful to realize precision matching for multi-source data(TM images, SPOT images, aerial photo) at different time. And its registration error can be decreased to the level of sub-pixel accuracy.

关 键 词：直方图匹配 特征点提取 影像匹配 局部纠正 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]