Estimation of 3D Geometry Using Multi-View and Structured Circular Light System

This paper proposes an approach to multiple view-projection systems planning for a 3D object reconstruction in a practical settings. A I/O system theoretic is adopted and, viewpoints and light sources which are located at different positions are respectively the output and the input. The system identification is defined by a relationship between input and output. Structured circular light patterns are generated by multiple light sources. The proposed approach may be applied to both multiple inputs(projectors)-viewpoints (MPV) or single inputviewpoint (SPV) systems. This contribution is chiefly for an approximate reconstruction sufficient to characterize / clarify a target object rather than a perfect 3D reconstruction. To that end, we propose a development of an efficient 3D system identification frameworks based around an efficient 3D camera system we have developed in the laboratory. We show that these are closely related to the geometric information of an object, and with minimal prior information, an efficient MPV system is achieved.