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：293-298
出版社：Copernicus Publications
摘要：This paper presents a hierarchical two-level method for reconstruction and detailed documentation of architectural scenes from photographs taken using digital cameras. The first level in the method involves creating a coarse model of the scene. Multiple images of an architectural scene are automatically matched for corresponding points and autocalibration is performed. Subsequently, a projective geometry based approach is adopted and bundle adjustment techniques are applied to create a reconstruction of the scene. The coarse model is thus generated with minimal user input, the results of which are presented in this paper. The second level takes the coarse model as the input and creates a detailed documentation of the architectural scene. Documentation consists of a 3D model and, associated with it, ancillary information about the architectural scene. The 3D model generated may be of a much higher geometric quality than the coarse model. Ancillary information is in the form of an information system containing attributes such as spatial reference, textual information about the structure, historical notes, conservation status and any other information of importance. The second level has two advantages – it avoids field reconnaissance and provides a means for automatic planning of camera positions. The advantages are offered by the concepts of virtual reconnaissance and Automatic Image Acquisition Planning (AIAP) respectively. As an extension, an integrated Architectural Heritage Information System (AHIS) is proposed. AHIS facilitates efficient management of both spatial and non-spatial ancillary information. The advantages of having a spatial database for architectural heritage will help conservation architects, government agencies and other interested parties to manage resources needed for preservation of built heritage effectively
