期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2008
卷号：XXXVII Part B4
页码：559-562
出版社：Copernicus Publications
摘要：This paper shows in theory and in practice how to implement a 3D reconstruction algorithm. It uses image sequence taken with a handheld camera as input to reconstruct a scene up to an unknown scale factor. The camera's motion and intrinsic parameters are all unknown. We especially address to apply SIFT algorithm to find distinctive features. Feature matching is done through a Euclidean-distance based nearest neighbor searching. The fundamental matrix is then estimated by existing correspondences and sequentially used to refine matching. By recovering project matrix from fundamental matrix we get projective reconstruction. And finally we demonstrate how the projective reconstruction can be successively upgraded to affine and Euclidean reconstruction
关键词：Image Sequence; Matching; Fundamental Matrix; Algorithm; 3D Reconstruction
