Proc. SPIE. 9890, Optical Micro- and Nanometrology VI
KEYWORDS: Microscopes, Mirrors, Optical sensors, 3D acquisition, Cameras, Microscopy, Optical microscopy, Optical microscopy, 3D modeling, Optical testing, Data acquisition, 3D metrology, Process control, Specular reflections, Digital micromirror devices, 3D scanning, 3D image processing, GPU based image processing, Structured light
In microsystems technology quality control of micro structured surfaces with different surface properties is playing an ever more important role. The process of quality control incorporates three-dimensional (3D) reconstruction of specularand diffusive reflecting technical surfaces. Due to the demand on high measurement accuracy and data acquisition rates, structured light optical microscopy has become a valuable solution to solve this problem providing high vertical and lateral resolution. However, 3D reconstruction of specular reflecting technical surfaces still remains a challenge to optical measurement principles. In this paper we present a measurement principle based on structured light optical microscopy which enables 3D reconstruction of specular- and diffusive reflecting technical surfaces. It is realized using two light paths of a stereo microscope equipped with different magnification levels. The right optical path of the stereo microscope is used to project structured light onto the object surface. The left optical path is used to capture the structured illuminated object surface with a camera. Structured light patterns are generated by a Digital Light Processing (DLP) device in combination with a high power Light Emitting Diode (LED). Structured light patterns are realized as a matrix of discrete light spots to illuminate defined areas on the object surface. The introduced measurement principle is based on multiple and parallel processed point measurements. Analysis of the measured Point Spread Function (PSF) by pattern recognition and model fitting algorithms enables the precise calculation of 3D coordinates. Using exemplary technical surfaces we demonstrate the successful application of our measurement principle.
In computer assisted quality control the three-dimensional reconstruction of technical surfaces is playing an ever more important role. Due to the demand on high measurement accuracy and data acquisition rates, structured light optical microscopy has become a valuable solution for the three-dimensional measurement of technical surfaces with high vertical and lateral resolution. However, the three-dimensional reconstruction of specular reflecting technical surfaces with very low surface-roughness and local slopes still remains a challenge to optical measurement principles. Furthermore the high data acquisition rates of current optical measurement systems depend on highly complex and expensive scanning-techniques making them impractical for inline quality control. In this paper we present a novel measurement principle based on a multi-pinhole structured light solution without moving parts which enables the threedimensional reconstruction of specular and diffuse reflecting technical surfaces. This measurement principle is based on multiple and parallel processed point-measurements. These point measurements are realized by spatially locating and analyzing the resulting Point Spread Function (PSF) in parallel for each point measurement. Analysis of the PSF is realized by pattern recognition and model-fitting algorithms accelerated by current Graphics-Processing-Unit (GPU) hardware to reach suitable measurement rates. Using the example of optical surfaces with very low surface-roughness we demonstrate the three-dimensional reconstruction of these surfaces by applying our measurement principle. Thereby we show that the resulting high measurement accuracy enables cost-efficient three-dimensional surface reconstruction suitable for inline quality control.