Absolute Phase Unwrapping with Deep Neural Network for Structured Light 3D Reconstruction

Phase-coding structured light is an important technique in 3D reconstruction. However, a great challenge is the wrapped phase that causes geometry ambiguity. Conventional unwrapping methods such as spatial and temporal approaches face the problem of error propagation and low efficiency. In this paper, we propose to solve the phase unwrapping problem with a deep neural network. To be specific, the phase unwrapping problem is cast to a semantic segmentation task, where the wrapped phase is the input and the fringe index for every pixel is the output. An encoder-decoder architecture, which is like U-net, is adopted as the network. We further propose a combined loss function by considering cross entropy loss, phase consistency loss and edge consistency loss. With 10000 artificially synthesized samples, the proposed method converges well. Experimental results demonstrate that the trained model well predicts fringe orders on both simulation data and real captured data. In addition, it unwrapps every pixel independently and avoids phase error propagation, and further achieves accurate 3D reconstruction.