An autostereoscopic three-dimensional (3D) metrology system for on-machine measurement of micro-structured surfaces

Abstract

Three-dimensional (3D) micro-structured surfaces possessing small-scale topologies are widely used in various fields. The geometric complexity and high accuracy requirement of these surfaces raise a lot of challenges regarding their measurement and characterization. The precision manufacture of 3D micro-structured surfaces demands ultra-precision machining and advanced error compensation technologies so as to ensure that the surfaces can be fabricated according to the stringent requirements. However, the error caused by the repositioning of the workpiece on the offline metrology system and the ultra-precision machine tool is substantial which has made the on-machine metrology system a good choice for the measurement of 3D micro-structured surfaces. In this thesis, an autostereoscopic 3D measurement method is presented and hence an autostereoscopic 3D metrology system is developed for on-machine measurement of micro-structured surfaces. The proposed method is divided into two stages including the information capture stage and 3D digital information reconstruction stage. During the first stage, the raw information of 3D surface profile is acquired by a series of 2D images of multiple perspectives with the aid of a micro lens array (MLA). These 2D images are slightly different as compared with each other and carry 3D information. More importantly, the proposed method constructs the quantitative relationship between the to-be-measured parameters and measured parameters acquired during the measuring process. The pixel information is used to carry the quantitative information in both the information capture process and 3D reconstruction process. Since the optical configuration of the two processes are symmetrical, the output is identical to the input 3D scene. The quantitative process of the autostereoscopic 3D surface measurement method allows the final output of a 3D digital model of a captured 3D object. Based on the autostereoscopic 3D surface measurement method, a 3D on-machine metrology system is developed. By incorporating a MLA into the imaging system, the system can capture 3D raw information of a workpiece through a single snapshot. For the 3D reconstruction process, a data processing method named direct extraction of disparity information (DEDI) is developed. The DEDI method is able to eliminate the defocus information with regard to certain depth of the raw data acquired during the information capture stage through a disparity pattern screening process and statistical analysis of the pixel information. This provides great convenience and high-precision 3D digital tomography-like reconstruction. The developed autostereoscopic 3D metrology system is further developed in order to perform on-machine measurement on ultra-precision machine tools. A series of experiments is conducted to measure the 3D micro-structured surfaces on a 4-axis ultra-precision machine tool, i.e. a Moore Nanoform 350FG. Repeated measuring is conducted on a sample to verify the performance of the autostereoscopic 3D metrology system. The results show that the system is able to achieve high reliability and accurate measurement. The experimental results of the proposed autostereoscopic on-machine 3D metrology system are further compared with the results generated by traditional offline measurement systems such as the ZYGO Nexview profiler system. It is interesting to note that the autostereoscopic 3D metrology system proposed and developed in this present study has the ability to provide sub-micrometer repeatability for on-machine measurement of micro-structured surfaces. On the whole, this research study contributes to an autostereoscopic 3D measurement method and hence the development of an autostereoscopic 3D metrology system which can perform on-machine measurement of 3D micro-structured surfaces. The technical feasibility of the proposed method and system is successfully verified through a series of on-machine measurements conducted on ultra-precision machine tools and compared with the results from a traditional offline measurement system. The outcome of the research not only contributes significantly to the advancement of measurement science and technology but also provides a turnkey solution for on-machine measurement of micro-structured surfaces.