人体三维扫描结构光解相位方法研究

Phase Retrieval of Structured Light on Three-Dimensional Body Scanning

在三维表面测绘中,基于格雷码和相移的结构光三维测量技术在临床医学中已被广泛应用.由于环境光、相机以及被测人体表面反光不均匀等因素的影响,相位展开时容易出现错误.针对上述问题本文提出一种算法,采用一幅没有条纹投影的目标图像和一幅经过条纹投影的图像共同完成图像分割,以期提高格雷码投影图像的分割准确率,消除目标表面纹理和环境光线的影响.再对格雷码图像和四步相移图像进行解码,得到绝对相位,并对绝对相位进行周期错位校正,最终获得理想的相位信息.实验系统由一个投影仪和平行双目摄像头组成,先后对仿真头部模型以及人的小臂进行三维扫描,生成的三维点云结果表明了该算法的可靠性.

In three dimensional （3D） surface mapping, the 3D measurement technology based on Gray code and phase shift structured light has been widely used in clinical medicine. Influence of surface re- flection of body, ambient light and camera results in mistakes during phase retrieving. In this paper, a new algorithm is proposed to solve the problem, which uses a target image and a Gray code image to real- ize image segmentation, in order to eliminate the influence of target surface textures and ambient light, and to improve the accuracy of Gray code projection image segmentation. And then the Gray code image and the four-step phase-shift image are decoded to obtain the absolute phase, which is corrected periodi- cally to obtain the ideal phase information. The experiment system consists of a projector and a parallel binocular camera. The 3D scanning of head model and human forearm is performed, and the generated 3 D point cloud proves the reliability of the algorithm.