中文摘要：

针对在实际应用中难以获得同源高分辨率遥感影像,利用同源高分辨率影像实现高精度、自动化的三维重建仍存在一定困难的现状,该文提出一种利用异源高分辨率遥感影像进行三维重建的方法.采用尺度不变特征变换算法对影像进行匹配以获取同名点,并使用随机抽样一致性算法剔除低精度匹配点;然后利用获取的同名点构建Delaunay三角网;接着根据有理函数反解模型和投影射线联合定位算法解算各个同名点的三维坐标;最后基于Delaunay三角网内插技术求取地物点的高程值,并基于OpenGL技术实现三维显示.试验结果表明,利用该方法可以实现异源高分辨率遥感影像的三维重建,精度上能够满足生产需求.

英文摘要：

It is difficult to obtain homologous high resolution remote sensing images in practical applications due to natural, technological and other factors. There are still some difficulties to achieve high-precision and automated three-dimensional reconstruction by using homologous high resolution images. Therefore, this paper proposed a three-dimensional reconstruction method by multi-source high resolution remote sensing images. Firstly, SIFT algorithm was used to match images in order to get corresponding points, and then RANSAC algorithm was used to embed the low-precision matching points; secondly, Delaunay triangulation was built by homogony points; then three-dimensional coordinates of each corresponding points were calculated based on inverse rational function model and projection rays stereo- scopic positioning algorithm; finally, elevation values of ground feature points were obtained by Delaunay triangulation interpolation technique and three-dimensional result was displayed by OpenGL technology. The experimental results showed that this method could achieve the three-dimensional reconstruction of multi-source high resolution remote sensing images, and the precision could meet practical requirements.