基于开源GIS数据的广域小重叠卫星影像定位模型精化  

Refinement of Geometric Model for Wide-Area and Small-Overlap Satellite Images Based on Open Source GIS Data

作　　者：王冬红[1] 刘军 黄石生 张蕾[1] 陈兴峰 WANG Donghong;LIU Jun;HUANG Shisheng;ZHANG Lei;CHEN Xingfeng(Beijing Institute of Tracking and Telecommunications Technology,Beijing 100094,China;Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China)

机构地区：[1]北京跟踪与通信技术研究所,北京100094 [2]中国科学院空天信息创新研究院,北京100094

出　　处：《上海航天（中英文）》2023年第2期60-66,共7页Aerospace Shanghai（Chinese&English）

基　　金：国家重点研发计划(2019YFE0126600)。

摘　　要：设计了一种适用于大区域、小重叠卫星影像有理多项式模型(RPC)定位精度优化的技术流程,基于开源数字正射影像(DOM)、数字高程模型(DEM)数据和RPC几何模型约束,通过优化尺度不变特征变换(SIFT)特征提取与匹配算法,实现了无像幅约束的控制点和连接点的快速稳健匹配,应用DEM支持下的RPC模型二维区域网方法实现了多景影像RPC模型精度的联合精化。GF-1卫星WFV影像的实验结果表明:在DEM支持下经过区域网平差后,RPC模型的定位精度可达到像素级水平。A technical process is designed to optimize the positioning accuracy of the rational polynomial camera(RPC)model for large-area and small-overlap satellite images.Based on the open source digital orthophoto map(DOM),digital elevation model(DEM)data,and RPC geometric model constraints,the scale-invariant feature transform(SIFT)feature extraction and matching algorithm is optimized to achieve the fast robust matching of control points and tie points free of the image size.The combined refinement of multiple RPC models is realized by DEMsupported two-dimensional bundle adjustment.The experimental results of the wide field of view(WFV)images of GF-1 satellite show that the positioning accuracy of the refined RPC model can reach the pixel level after the bundle adjustment supported by the DEM.

关 键 词：有理多项式模型(RPC)模型 区域网平差 尺度不变特征变换(SIFT) 随机采样一致性 特征匹配 

分 类 号：P236[天文地球—摄影测量与遥感]