Abstract
An automatic surface registration method is developed based on Regional Curvature Map, an automatic surface registration method is developed. Matched point pairs between the two surfaces are established by matching RCMs, and the matched point pairs that satisfy the geometric constraints are selected to calculate the initial alignment transformation, which is fine-tuned by iterative closest point algorithm and verified. Experimental result shows that the proposed method can register two overlapping scanned surfaces effectively. It is efficient, robust to noise, and has a high successful rate.
| Original language | English |
|---|---|
| Pages (from-to) | 245-248 |
| Number of pages | 4 |
| Journal | Guangdianzi Jiguang/Journal of Optoelectronics Laser |
| Volume | 21 |
| Issue number | 2 |
| State | Published - Feb 2010 |
Keywords
- Curvature
- Geometric constraint
- Matched point pairs
- Point descriptor
- Surface registration
Fingerprint
Dive into the research topics of 'Automatic surface registration based on regional curvature map'. Together they form a unique fingerprint.Cite this
- APA
- Author
- BIBTEX
- Harvard
- Standard
- RIS
- Vancouver