Research on reconstruction of surface morphology by optical scanning using a Dammann grating

微光学是研究特征尺度为微米级的光学元件和基于微光学元件的集成光学系统的现代光学分支。作为微光学的重要分支，二元光学是基于光波衍射理论发展起来的、光学与微电子技术相互渗透、交叉形成的前沿学科。随着二元光学和微加工工艺的发展和创新，传统的大尺寸光学元件逐渐转为微型化、集成化，实现了大规模制造。微光学元件是制造小型光电子系统的关键元件，它具有体积小、质量轻、造价低等优点，并且能够实现普通光学元件难以实现的微小阵列、集成、成像和波面转换等功能。二元光学元件的典型代表之一是Dammann光栅，它具有效率高、体积小、易于复制等优点。在光学三维测量领域，使用Dammann光栅可以极大的提高光学扫描效率以及具有很高的信噪比。此外，由于Dammann光栅的各级衍射斑的相对距离严格固定，故而降低了单点扫描时转镜的转动误差。本文旨在研究Dammann光栅在光学三维测量领域的应用。论文主要内容如下： 1. 研究了基于Dammann光栅的物体表面形貌测量的机理和过程，提出了用Dammann光栅增加光学点扫描测量物体形貌的扫描效率以及保持高的信噪比。利用Dammann光栅的分光特性，单点光源被分束成多个光强一致的衍射光斑，衍射光斑入射到物体表面，多个点被物体表面的高度信息调制，从而实现了多点同时扫描，避免了单点扫描的扫描效率的限制。同时，该方法避免了条纹投影所需要的均匀照明环境，也不存在条纹混叠造成的分辨率极限。尤其适用于远距离、大尺寸物体的三维表面形貌的获取。在三维工件的立体数字化建模，飞机上对地面建筑、地形等三维数字化信息获取，卫星上多点激光扫描获得月球表面三维信息等众多领域，具有广泛地应用前景。完成了相关代码的编写，进行了相应的实验。 2. 讨论了计算机视觉中摄像机的标定过程。论文研究了针孔摄像机模型的标定过程，建立了摄像机坐标系和世界坐标系的坐标对应关系，给出了转换的旋转矩阵和平移矩阵，确定了摄像机的内外参数。在此基础上，重点研究了基于三角法的单目视觉的摄像机定位方法，它是针对本扫描装置简化的标定过程。该标定方法实现了对实验目标物体表面准确的三维测量，结果证明了该方法的鲁棒性。As a major branch of modern optics, micro-optics istudies micro-scale optical elements and micro-scale integrated optical systems. Developments of binary-optics and micro-machining technologies make it possible to reduce the large size of traditional optical components into miniaturization and integration of optical devices gradually. Micro-optical elements are the key components of small optoelectronic systems, for its small size, light weight and low cost advantages. And furthermore, it can easily achieve the small array, integration, imaging and wave front conversion function, where traditional optical components are difficult to realize. Dammann grating is a typical binary optical element with the advantage of high diffraction efficiency, small size, and easy replication. In optical three-dimensional measurement, the use of Dammann grating can greatly improve the efficiency of optical scanning and can achieve very high spatial resolution. In addition, it can achieve a high contrast ratio of the signal to the background noise in the diffractive pattern and each diffraction spot has equal distance to its neighbors, and therefore Dammann grating minimizes the rotation error of galvanometer used in single-point optical scanning. This dissertation aims to study the application of Dammann grating in optical three-dimensional measurement field. The main contents are as follows: 1. Reconstruction of three-dimensional surface morphology by optical spot-scanning using a Dammann grating is presented. We proposed a new measurement set in the process of surface morphology by optical spot-scanning with a Dammann grating to increase the scanning efficiency and maintain the high spatial resolution. A one-dimensional (1D) Dammann grating is capable of splitting more than 80% of incident power into desired diffractive orders. Compared with the nonuniform spots generated by normal amplitude grating, the orders have almost equal intensity. Besides, it achieves a high contrast ratio of the signal to the background noise in the diffractive pattern and each diffraction spot has equal distance to its neighbors. It avoids the requirement for fringe projection, and there is no resolution limit caused by stripe-aliasing. A variety of applications of this 3D shape measurement will include measuring the surface profile of a large scale object in a long distance, such as the large object on the ground, the lunar surface profile and so on. 2. Camera calibration process in computer vision was studied. Calibration of pin-hole camera model was used. We have established the coordinate mapping relationship between camera coordinate system and world coordinate system. We derived the rotation matrix and translation matrix of conversion, and furthermore, we found the internal and external camera parameters. Then we focused on camera positioning of monocular vision based on triangulation method. It is a simplified calibration for our three-dimensional measuring set. This calibration method realized accurate reconstruction of target surface topology, and the results demonstrated the robustness of this method