Editing Point cloud (section)

== Alignment and registration ==
{{Main|Point set registration}}
When scanning a scene in real world using [[Lidar]], the captured point clouds contain snippets of the scene, which requires alignment to generate a full map of the scanned environment.

Point clouds are often aligned with 3D models or with other point clouds, a process termed [[point set registration]].

The [[Iterative closest point|Iterative closest point (ICP) algorithm]] can be used to align two point clouds that have an overlap between them, and are separated by a [[Rigid transformation|rigid transform]].<ref>{{Cite web |title=Continuous ICP (CICP) |url=https://www.cs.cmu.edu/~halismai/cicp/ |access-date=2024-06-26 |website=www.cs.cmu.edu}}</ref> Point clouds with elastic transforms can also be aligned by using a non-rigid variant of the ICP (NICP).<ref>{{Cite journal |last1=Li |first1=Hao |last2=Sumner |first2=Robert W. |last3=Pauly |first3=Mark |date=July 2008 |title=Global Correspondence Optimization for Non-Rigid Registration of Depth Scans |url=https://onlinelibrary.wiley.com/doi/10.1111/j.1467-8659.2008.01282.x |journal=Computer Graphics Forum |language=en |volume=27 |issue=5 |pages=1421–1430 |doi=10.1111/j.1467-8659.2008.01282.x |issn=0167-7055}}</ref> With advancements in [[machine learning]] in recent years, point cloud registration may also be done using [[End-to-end principle|end-to-end]] [[Neural network (machine learning)|neural networks]].<ref>{{Cite journal |last1=Lu |first1=Weixin |last2=Wan |first2=Guowei |last3=Zhou |first3=Yao |last4=Fu |first4=Xiangyu |last5=Yuan |first5=Pengfei |last6=Song |first6=Shiyu |date=2019 |title=DeepVCP: An End-to-End Deep Neural Network for Point Cloud Registration |url=https://openaccess.thecvf.com/content_ICCV_2019/html/Lu_DeepVCP_An_End-to-End_Deep_Neural_Network_for_Point_Cloud_Registration_ICCV_2019_paper.html |pages=12–21}}</ref>

For industrial metrology or inspection using [[industrial computed tomography]], the point cloud of a manufactured part can be aligned to an existing model and compared to check for differences. [[GD&T|Geometric dimensions and tolerances]] can also be extracted directly from the point cloud.