题目：Smart Photogrammetry for 3D Shape Measurement

时间：2019年10月22日 10:00-11:00

地点：机械与动力工程学院 F103会议室

邀请人：冯晓冰 助理教授（制造技术与装备自动化研究所）

Biography

Dr. Samanta Piano (Faculty of Engineering, Manufacturing Metrology Team) is an Assistant Professor at the University of Nottingham. Her current research interest concerns the development of innovative and unconventional optical techniques and 3D probing systems for high-precision coordinate metrology to be used in industrial applications. Samanta received her PhD in Physics in 2007, researching on high temperature superconductors and nanofabrication of strongly ferromagnetic Josephson junctions, useful for applications to spintronics and quantum computing. Samanta's career in Nottingham has been supported by a Marie Curie Intra-European Fellowship (2009-2011) and a Nottingham Advance Research Fellowship. She has published in international peer reviewed journals (attracting about 500 citations; source: Scopus), including three in the prestigious Physical Review Letters, one in Applied Physics Letters, and three refereed book chapters. She is currently co-supervising six PhD students, working on novel optical measurement techniques, complex mechanism error modelling and close range photogrammetry. She has recently awarded a Hermes Fellowship to work on application of machine-learning and CAD model recognition techniques to accelerate and optimise the performance of the photogrammetry system.

Abstract

When new manufacturing technologies and associated component designs emerge, suitable metrology techniques for process development, optimisation and inspection can lag behind. This is the case with additive manufacturing (AM), particularly regarding the complex geometries of component designs, which are correspondingly difficult to measure.  An application-specific metrology solution must achieve an optimal trade-off between factors such as resolution, measurement range, accuracy, repeatability, cycle time and costs. At the scale appropriate to AM (10 mm to 1000 mm), photogrammetry presents itself as an attractive, low-cost candidate technology, able to acquire high accuracy 3D form data. The research within the MMT is developing a camera-based photogrammetry system capable of addressing AM applications. By acquiring multiple 2D component images, the system software reconstructs a 3D point cloud which represents the component's surface geometry. Current research is devoted to deliver software able to perform calibration, set-up, data capture and post-process at high speed. Information-rich metrology together with artificial intelligence and CAD model recognition are implemented into the measurement process, allowing the system performance to be optimised by determining camera and object orientation, accelerating feature-matching, and positioning the cameras during data capture.