Classification and Characterization of Image Acquisition for 3D Scene Visualization and Reconstruction Applications

Abstract

This paper discusses the techniques of image acquisition for 3D scene visualization and reconstruction applications (3DSVR). The existing image acquisition approaches in 3DSVR applications are briefly reviewed. There are still lacks of the studies about what principles are essential in the design and how we can characterize the limitations of an image acquisition model in a formal way. This paper addresses some of the main characteristics of existing image acquisition approaches, summarized through a classification scheme and illustrated with many examples. The results of the classification lead to general characterizations in establishing the notions (basic components) for design, analysis and assessment of image acquisition models. The notions introduced include: focal set, receptor set, reflector set etc. The definitions of the notions are given and supported with various examples (e.g. concentric panoramas, cataoptrical panoramas). The observations, important issues, and future directions from this study are also provided.