3D Data Fusion for Multimodal Inspection of Mechanical Parts

Data fusion is a key step in multimodal inspection of mechanical parts. It combines complementary data from different inspection methods, or “modalities”, to improve inspection accuracy. Data fusion has broad applications in the disciplines of the military, medicine, and geography. It has attracted increasing attention from industry in recent years. However, its industrial application has been limited mainly to 2D (2-Dimensional) domains. The research focus of this work is to develop a data fusion system to automatically fuse and extract the most accurate information from each modality in the domain of 3D. The main process includes automatically filtering and interpolating CMM data, registering CMM data with CT data, detecting the CT slices to generate interpolated CMM constraints for iterative CT reconstruction, and replacing exterior CT data with interpolated CMM data. This technology enables us to achieve high accuracy for external structures and provide constraints for CT reconstruction to improve the accuracy for internal boundary data as well.Copyright © 2004 by ASME