Qian Jiaming, Feng Shijie, Li Yixuan, Tao Tianyang, Han Jing, Chen Qian, Zuo Chao
Opt Lett. 2020 Apr 1;45(7):1842-1845. doi: 10.1364/OL.388994.
Recovering the high-resolution three-dimensional (3D) surface of an object from a single frame image has been the ultimate goal long pursued in fringe projection profilometry (FPP). The color fringe projection method is one of the technologies with the most potential towards such a goal due to its three-channel multiplexing properties. However, the associated color imbalance, crosstalk problems, and compromised coding strategy remain major obstacles to overcome. Inspired by recent successes of deep learning for FPP, we propose a single-shot absolute 3D shape measurement with deep-learning-based color FPP. Through "learning" on extensive data sets, the properly trained neural network can "predict" the high-resolution, motion-artifact-free, crosstalk-free absolute phase directly from one single color fringe image. Compared with the traditional approach, our method allows for more accurate phase retrieval and more robust phase unwrapping. Experimental results demonstrate that the proposed approach can provide high-accuracy single-frame absolute 3D shape measurement for complicated objects.
从单帧图像中恢复物体的高分辨率三维（3D）表面一直是条纹投影轮廓术（FPP）长期追求的最终目标。彩色条纹投影方法因其三通道复用特性，是最有潜力实现这一目标的技术之一。然而，相关的颜色不平衡、串扰问题以及不完善的编码策略仍然是需要克服的主要障碍。受深度学习在FPP领域近期成功的启发，我们提出了一种基于深度学习的彩色FPP单次绝对3D形状测量方法。通过在大量数据集上“学习”，经过适当训练的神经网络可以直接从一幅彩色条纹图像中“预测”出高分辨率、无运动伪影、无串扰的绝对相位。与传统方法相比，我们的方法能够实现更精确的相位检索和更稳健的相位展开。实验结果表明，所提出的方法能够为复杂物体提供高精度的单帧绝对3D形状测量。
