Application of Thin-shell Finite Elements in Range Image Registration and Surface Smoothing
碩士 === 國立中興大學 === 機械工程學系所 === 96 === Every single view profile information obtained by measurement system has it’s own measurement error. Though executing the range image registration in all view could reduce the errors between every view to the best of it’s possible, there are calculational errors...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | zh-TW |
| Online Access: | http://ndltd.ncl.edu.tw/handle/93682570707482195661 |
| Summary: | 碩士 === 國立中興大學 === 機械工程學系所 === 96 === Every single view profile information obtained by measurement system has it’s own measurement error. Though executing the range image registration in all view could reduce the errors between every view to the best of it’s possible, there are calculational errors made form shape integrated. It causes the erroneous faults occur between every range image and any other range images which overlapping.
This paper describes using triangular thin-shell finite elements to mesh every singe view range image. For seeking the corresponding relationship between every single view profile, we determine the overlaps with other range images to find out the all nodes’ corresponding points. After that we use the virtual springs which original lengths are zero to connect them, and use the spring’s restoring force to minimize the whole error of range images and reduce the influence about object’s surface profile to get the whole contours of component.
|
|---|