Partial Alignment of Data Sets Based on Fast Intrinsic Feature Match
- 815 Downloads
Point Feature Histograms (PFH) is a statistic and geometric invariant descriptor that has been widely used in shape analysis. Current PFH based feature extraction methods are highly affected by the time scale and become less effective for unbalanced cases, which limits their performance. In this paper, we focus on finding a framework for partial registration by an adaptive partition of point set algorithm. Firstly, we propose an adaptive partition method base on PFH coding. Secondly, we conduct a series of fast parallel implementations for efficiency. Thirdly, we plug in the PFH based partition method and trimmed strategy to our modified iterative closest point method. Experiments demonstrate that our algorithms are robust and stable.
KeywordsPartial registration PFH coding ICP Trimmed method
This work was supported by the National Natural Science Foundation of China under Grant No. 11771276.
- 13.Johnson, A.E., Hebert, M.: Surface registration by matching oriented points. In: Proceedings of the International Conference on Recent Advances in 3-D Digital Imaging and Modeling, NRC 1997, p. 121. IEEE Computer Society (1997)Google Scholar
- 17.Zha, H., Ikuta, M., Hasegawa, T.: Registration of range images with different scanning resolutions. In: Proceedings of the IEEE International Conference on System, Man, Cybernetics, Nashville, Tennessee, USA, pp. 1495–1500 (2000)Google Scholar