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：2010
卷号：XXXVIII - Part 5
页码：55-61
出版社：Copernicus Publications
摘要：Automation in 3D reconstruction involves mainly two tasks: (i) image orientation and (ii) 3D scene reconstruction. Nowadays, automation in (i) can be considered solved for aerial photogrammetric projects, but a general and commercial solution is still pending in case of close-range image blocks. Here, the complexity and diversity of the network geometry make the automatic identification of the image correspondences a challenging task. In this paper we present a methodology to automatically orient a set of targetless close-range images taken with a calibrated camera and the following creation of 3D models with a multi-image matching algorithm. Several real cases, comprehending free-form objects (e.g. bas-relieves, decorations, etc.), building facades and images taken with UAVs are reported and discussed. Lastly, some preliminary results on a methodology for markerless camera calibration are presented
