Deflectometry for technical surfaces (DOTS)

在许多生产过程中，质量控制都需要一种快速、稳健、非接触、高精度的三维形状测量技术。鉴于这些要求，光学、非干涉技术显得非常合适。然而，这些技术的适用性和准确性在很大程度上取决于所考虑的对象的表面属性。对于表面粗糙度尤其如此，对于不同的技术表面可能会有很大的不同。单个技术对象通常具有一定程度的光学平滑(镜面或镜面)区域以及粗糙区域，这使得高精度、可靠的测量具有极大的挑战性。相位测量偏转测量(PMD)是一种几何光学计量技术，主要用于镜面对象的3D形状测量。基于PMD的系统原则上非常适合于技术表面的表征，因为与触觉形式测量系统相比，它们更小、更便宜、更快。然而，偏振模色散的理论框架是以完全镜面为前提的，忽略了物体表面的粗糙度和波纹度。技术对象表面通常不能满足这些要求，并导致额外的测量误差到目前为止还没有得到充分的表征，因此，本项目的目标是开发技术表面上真实的PMD测量表面模型，能够预测与表面相关的统计数据，并减少系统测量误差。要完成的子目标是：i)对测量对象的形状、PMD设置的几何形状和相位测量误差之间的关系进行建模，由样品对象的测试测量和测量模拟来支持；ii)对将相位测量误差传播到表面形状、梯度和曲率的导出量的建模。用于实现子目标i)的模型是双向反射比分布函数。求解该函数的方法有Phong-Procedure方法和蒙特卡罗方法。子目标II)通过最小二乘积分和径向基函数积分得到，利用所建立的模型，可以预测系统的PMD测量误差并确定其原因，从而可以对其进行校正。对于不同类型的物体表面，也可以定量地确定统计不确定度。这使得PMD形式的测量第一次有可能：i)系统地调整PMD设置和测量程序以适应被测物体表面；ii)通过校正与物体表面相关的系统误差来提高PMD测量精度；以及iii)计算与物体表面相关的空间分辨的统计测量不确定度图。在项目结束时，应提供适用于技术表面的优化的偏转测量过程，以弥补工业质量控制的显著差距。

Quality control requires, for many production processes, a fast, robust, non contact and high precision measurement technique for 3D form measurement. In view of these requirements, optical, non-interferometric techniques appear very suitable. The applicability and accuracy of these techniques, however, strongly depends on the surface properties of the objects under consideration. This is especially true for the surface roughness which may be very different for various technical surfaces. Individual technical objects often have to some extent optically smooth (mirror like or specular) areas as well as rough areas, making high-precision, reliable measurements extremely challenging.Phase Measuring Deflectometry (PMD) is a geometric-optical metrology technique primarily used for 3D form measurement of mainly specular objects. Systems based on PMD are in principle well suited for the characterization of technical surfaces, as they are smaller, cheaper and faster compared to tactile form measurement systems. However, the theoretical framework of PMD presupposes perfectly specular surfaces, neglecting object surface roughness and waviness. Technical object surfaces usually do not fulfil these demands and lead to additional measurement errors not adequately characterized up to now.Thus, the goal of this project is to develop realistic surface models for PMD measurements on technical surfaces that can predict the surface-related statistic and decrease systematic measurement errors. The sub-goals to be accomplished are: i) Modelling of the relationship between shape of the measurement object, geometry of the PMD setup and phase measurement errors, supported by test measurements of sample objects and simulation of measurements; ii) Modelling of the propagation of phase measurement errors into the derived quantities of surface forms, gradients and curvatures. The model used for achievement of sub-goal i) is the Bidirectional Reflectance Distribution Function. The approaches to this function are the Phong-procedure and Monte-Carlo methods. Sub-goal ii) is obtained by employing Least Squares Integration as well as Radial Basis Function Integration.With the models developed, systematic PMD measurement errors can be predicted and their reasons identified, and thus they can be corrected. Also the statistic uncertainties can be quantitatively determined for different types of object surfaces. This makes it possible for the first time for PMD form measurements to i) systematically adapt the PMD setup and the measurement procedures to the object surface to be measured; ii) increase the PMD measurement accuracy by correcting object surface-related systematic errors; and iii) calculate spatially resolved statistic measurement uncertainty maps related to the object surface.At the end of the project an optimized deflectometric measurement process suitable for technical surfaces shall be available in order to close a significant gap for industrial quality control.