Structured-light-based highly dense and robust 3D reconstruction.

In this paper, a structured-light-based highly dense and robust 3D reconstruction method is proposed by combining a Gray code and region-shifting pattern. The region-shifting pattern is transformed to the trapezoidal and triangle wave shifting pattern by combining all frames of the region-shifting pattern, and then the boundary of the trapezoidal wave shifting pattern and the peak and phase of the triangle wave shifting pattern are estimated. Through this technique, the spatial resolution is increased about three times. Consequently, the 3D points are reconstructed with a resolution much higher than a camera image resolution. Moreover, as the proposed method measures the boundary and the peak with all frames, it increases the signal-to-noise ratio and is more robust than the conventional methods that use only one or two frames to detect them.