3D reconstruction methods for digital preservation of cultural heritage: A survey

摘要

3D reconstruction, refers to capturing and reproducing the shape and appearance of an arbitrary object or scene given depth and color information. This is a broad research area within the computer vision field involving many stages and still open problems. The digital preservation of cultural heritage is a specially challenging application of 3D reconstruction. Cultural heritage objects and sites greatly differ from each other and a maximized fidelity of the 3D reconstruction is a core requirement. The literature on this topic has substantially increased in the past years, mostly due to the variety of scenarios and the development of new depth sensing devices as well as techniques able to deal with this issue. In our search to develop a complete 3D reconstruction pipeline, we have comprehensively studied techniques related to this topic and divided the 3D digitization process in four major overviews: image acquisition, view registration, mesh integration and texture generation. We present the state-of-the-art approaches and challenges of each stage.