2. Registration
要解决的问题:Reconstruction from scans
多个视点的扫描,每个视点得到分片3D数据,如何将这些分片数据合并成一个整体数据?
各个角度采集到的特体是碎片的,怎么拼起来?
Planing 怎么扫不同角度得到完整数据
Pairwise Registration
目标:
Align a source model χ onto a target model y,find a transformation T χ that brings χ into alignment with y
挑战:
- How do we measure the quality of the alignment?
- What transformations are acceptable?
基本方法:Registration as energy minimization-
energy包含两部分:Alignment Error和Transformation Error,分别对应解决上面的两个挑战。
[46:21] Ematchi :匹配误差, Eprior:先验知识
Alignment Error
Prior Error – 1 (Rigid objects)
Eprior(Z,R,t)=n∑i=1||Zi−(RXi+t)||22
Prior Error – 2 (Elastic objects)
弹性物体,存在少量变形
Eprior(Z,(Ri))=n∑i=1∑j∈Ni||(Zj−Zi)−Ri(Xj−Xi)||22
Prior Error – 3 (Articulated objects)
分段刚性
Iterative Closest Point (ICP) Algorithm
[Besl+92]
根据对应点求变换关系(R.t)
根据变换关系就对应点
本文出自CaterpillarStudyGroup,转载请注明出处。 https://caterpillarstudygroup.github.io/GAMES102_mdbook/