Integrated Measurement System for Rapid Shape Measurement, Model Verification and NDE
Final rept. 1 Sep 1996-31 Aug 2000
SOUTH CAROLINA UNIV COLUMBIA DEPT OF MECHANICAL ENGINEERING
Pagination or Media Count:
Extensive studies of shape measurement systems have been completed. Both numerical simulations and controlled experiments have been performed. Results have shown that, for fringe patterns having sensor plane fringe spacing from 3 to 30 pixels, the resulting phase error can be reduced to less than 0.02 radians by controlling the spatial frequency of the signal and noise, as well as the contrast in the pattern. Studies of data integration methodologies have also been completed. Results demonstrate that the procedure developed by the authors is both robust and accurate when registering and integrating multiple, overlapped data sets. The effects of key variables in the registration process have been quantified through numerical simulations. Results indicate that when the feature size divided by the overlapped region size, alpha fsls - 0.5, and the mesh size divided by the feature size, beta msfs - 0.25, the algorithm can accurately and efficiently register and integrate multiple data sets with a relative error, sigma rams - 10. Finally, studies of automated registration processes have been performed. Local differential geometry properties such as gaussian curvature, K, and mean curvature, H, along with their signs have been used to successfully automate the process of integration of separate shape measurement data sets.
- Numerical Mathematics
- Theoretical Mathematics
- Manufacturing and Industrial Engineering and Control of Production Systems