True Three-Dimensional Reconstruction of Multi-View Satellite Stereo Images of Alpine Valley
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 topo-graphic mapping,early identification and monitoring of geological disasters,and analysis of glacier evolution in high-altitude areas.The study re-sults can provide reference for true 3D reconstruction of satellite images.
alpine valleysatellite stereo image pairsrational function modelimage matchingtrue 3D reconstruction
国家科技期刊平台
NSTL
万方数据
