基于同名点对的小面元影像微分纠正及融合方法  

Tiny Facet Image Differential Rectification and Fusion Approach based on Corresponding Image Points

作　　者：莫华[1] 秦志远 黄华文 

机构地区：[1]信息工程大学测绘学院研究生 [2]不详

出　　处：《测绘科学与工程》2006年第3期27-32,共6页Geomatics Science and Engineering

摘　　要：多源遥感影像数据融合是提高海量遥感影像数据利用率，获取更多决策辅助信息的有效手段之一。目前一些商业遥感图像处理软件在进行多源遥感影像融合时，存在效率低、精度难以达到实用要求等问题。本文提卅了基于图像线特征的遥感影像精确配准、纠正、融合新方法，该方法运用线特征算法提取图像边缘特征点作为控制点，然后通过全自动图像匹配获取密集同名点对，由这些同名点对构成密集三角网（小面元）实现影像精确配准；最后进行影像纠正融合处理。大量实践表明，该方法是一种高效实用、高精度的遥感影像配准、纠正、融合方法，能大大改善山区等特殊地区的配准精度。Multi - sensor Remote Sensing Image Data Fusion （MRsIDF） is one of efficient approaches to improeve the profitability of enormous Remote Sensing （RS） image data and obtain more information for making decision. At present, some commercial software for RS image are inefficient and can not meet the requirements of application in the process of MRsIDF. In this paper, a new method based on the line feature for RS image accurate registration, rectification and fusion is presented. Line feature algorithm is used to extract the edge feature points of image as control points. Corresponding image points which constitute triangulation network to achieve accurate image matching are acquired by automatic image matching. Finally , the image rectification and the fusion processing are executed. Lots of practices indicate that this method is an efficient, applicable and precise means for image registration, rectification and fusion. It can improve the matching accuracy greatly especially for mountain region.

关 键 词：线特征 图像匹配 小面元 图像融合 

分 类 号：P237.3[天文地球—摄影测量与遥感]