Acta Optica Sinica, Volume. 38, Issue 12, 1215001(2018)
Cross-Scale Registration Method Based on Fractal Dimension Characterization
Fig. 1. Cross-scale data obtained by different measurement configures
Fig. 2. Multilevel wavelet transform space
Fig. 3. Process of two-dimensional discrete wavelet transform
Fig. 4. Flowchart of ICP algorithm
Fig. 5. Surface texture of cutting tool
Fig. 6. Three-dimensional topographical data at different magnifications. (a) 20×; (b) 50×; (c) 100×
Fig. 7. (a) Raw data; three-dimensional data after DWT at (b) level 3 and (c) level 5
Fig. 8. Two-dimensional contour of different levels at the same position of DWT data. (a) Level 0; (b) level 1; (c) level 2; (d) level 3; (e) level 4; (f) level 5
Fig. 9. Logarithmic graph of (a) raw data at magnification of 100× and data after DWT at (b) level 3 and (c) level 5
Fig. 10. Logarithmic graph of raw data at different magnifications. (a) 20×; (b) 50×; (c) 100×
Fig. 11. (a) Registration results of scale-approximated data of level 3 wavelet approximation of 100× small scale data and 20× data; (b) apply the transform matrix on raw data; (c)(d) corresponding local magnification diagram
Fig. 12. (a) Two-dimensional contour of registration result of scale-approximated data; (b) two-dimensional contour of raw data after applying transform matrix
|
|
Get Citation
Copy Citation Text
Chengwei Mo, Haihua Cui, Xiaosheng Cheng, Haibin Yao. Cross-Scale Registration Method Based on Fractal Dimension Characterization[J]. Acta Optica Sinica, 2018, 38(12): 1215001
Category: Machine Vision
Received: May. 24, 2018
Accepted: Jul. 12, 2018
Published Online: May. 10, 2019
The Author Email: