高山峡谷多视卫星立体像对真三维重建  

True Three-Dimensional Reconstruction of Multi-View Satellite Stereo Images of Alpine Valley

作　　者：杨刚[1] 吕宝雄[1] 贾玮 王明[1] 杨振胤 李贾亮 刘康 YANG Gang;LYU Baoxiong;JIA Wei;WANG Ming;YANG Zhenyin;LI Jialiang;LIU Kang(PowerChina Northwest Engineering Corporation Limited,Xi'an 710065,China;College of Geomatics,Xi'an University of Science and Technology,Xi'an 710054,China)

机构地区：[1]中国电建集团西北勘测设计研究院有限公司,西安710065 [2]西安科技大学测绘与科学技术学院,西安710054

出　　处：《西北水电》2024年第4期11-17,共7页Northwest Hydropower

基　　金：国家自然科学基金青年基金项目(41904036).

摘　　要：为解决大面积艰险条件下高山峡谷真三维重建难题,提高卫星影像空间及光谱分辨率,以中国西藏某高山峡谷流域Pleiades卫星立体像对为切入点,探讨了RPC轨道参数校正、SIFT算子优化密集匹配等卫星影像三维重建关键技术,构建无控定位条件下全色、多光谱及融合状态下真三维重建模型。结果表明:3种方案下多视卫星立体像对匹配点误差均小于1个像素,全色方案匹配点精度达到0.67个像素,丰富了1∶5千、1∶1万地形图测绘、地质灾害早期识别及监测、高海拔地区冰川演变分析手段及方法。研究成果可为卫星影像真三维重建提供借鉴。In order to solve the problem of true 3D reconstruction of large-scale alpine valley under difficult conditions and improve the spatial and spectral resolution of satellite images,the Pleiades satellite stereo image pairs of an alpine valley basin in Tibet are used as the starting point to explore the key technologies of satellite image 3D reconstruction,such as RPC orbit parameter correction and SIFT operator optimization dense matching,and to build a true 3D reconstruction model under panchromatic,multispectral and fusion conditions without controlled positioning.The results show that the matching point errors of multi-view satellite stereo image pairs under the three schemes are all less than 1 pixel,and the matching point accuracy of the panchromatic scheme reaches 0.67 pixels,which enriches the means and methods of 1∶5000 and 1∶10000 topographic mapping,early identification and monitoring of geological disasters,and analysis of glacier evolution in high-altitude areas.The study results can provide reference for true 3D reconstruction of satellite images.

关 键 词：高山峡谷 卫星立体像对 有理函数模型 影像匹配 真三维重建 

分 类 号：P237[天文地球—摄影测量与遥感]