基于表面反射特性建模的高鲁棒性微相位编码结构光三维扫描技术研究

非朗伯表面具有镜面反射、全局反射及局部多次反射等特性，是结构光三维重建领域的研究难点，现有的结构光三维重建方法在应对非朗伯表面时存在鲁棒性不足的问题。本项目拟提出一种高鲁棒性的微相位编码结构光三维扫描技术，将光度学中的表面反射建模理论引入至结构光三维重建研究中，通过对非朗伯表面的反射特性进行建模，建立表面反射率与高频正弦条纹亮度值之间的函数关系，实现对结构光编码图像的调整，进而实现非朗伯表面的高鲁棒性、高精度三维重建。本项目的研究内容主要包括：1）物体表面反射特性估计与建模；2）基于高频正弦条纹投射的局部多次反射特性分析；3）基于反射模型估计的结构光编码图像调整；4）微相位编码结构光的编解码策略。通过对上述内容的研究，以期实现在不经消反光处理条件下非朗伯表面的直接高精度三维重建。该项研究有助于提高结构光三维扫描技术的实用性，并大大地拓展其应用范围，因而具有重要的研究价值和应用前景。

Non-Lambert surface has the features of specular reflection, global reflection and local inter-reflection. Its measurement is a difficult research topic in structured light sensing (SLS)-based 3D reconstruction field. The reason is that the existing structured light techniques are lack of robustness while measuring non-Lambert surface. To solve this problem, in this proposal, a micro-phase coded structured light method with high robustness is investigated. The surface reflection modeling method in photometry is introduced into SLS-based 3D reconstruction to estimate the surface reflectivity and reflection property. By establishing the correlation function between the surface reflectivity and intensity of the projected high-frequency sinusoidal fringe pattern, the projected structured light patterns are adjusted. Thus, 3D reconstruction with high robustness and precision is achieved. This proposal will focus on the following key research problems: 1) Estimation of the surface reflectivity and reflection property; 2) Analysis of local inter-reflection property based on the projection of high-frequency sinusoidal fringe pattern; 3) Adjustment of the coded structured light pattern based on the estimation of the surface reflection model; 4) Investigate the encoding and decoding strategies of micro-phase coded structured light. By solving above key problems, this proposal aims to realize direct and precise 3D reconstruction of non-Lambert surface without the onerous surface painting procedure, so as to further improve the 3D reconstruction quality. With these research contributions, the practicability of SLS 3D scanning techniques can be improved distinctly, and its application fields can also be greatly extended.