A low-cost 3D reconstruction system using a single-shot projection of a pattern matrix

Abstract

The purpose of this work is to present the main features of a structured light system and some experimental results. The system, named SingleShot3DSL, is based on a single off-the-shelf digital camera and a pattern projector. The mathematical model for 3D reconstruction is based on the collinearity equations. A pattern codification strategy was developed to allow fully automatic pattern recognition. A calibration methodology enables the determination of the direction vector of each pattern and the coordinates of the perspective centre of the pattern projector. The calibration processes are carried out with the acquisition of several images of a plane taken from different distances and orientations. Several processes were combined to provide a reliable solution for pattern location. In order to assess the accuracy and potential of the methodology a prototype was built, integrating a pattern projector and a digital camera in a single mount. Experiments using reconstructed surfaces with real data indicated that a relative accuracy of 0 center dot 2 mm in depth could be achieved.