3D Registration in 30 Years: A Survey

TL;DR

This survey comprehensively tracks 30 years of 3D point cloud registration, detailing pairwise coarse/fine, multi-view, cross-scale, cross-source, color, and multi-instance problems. It presents a systematic taxonomy that spans geometric and learning-based methods, including correspondence-based and correspondence-free approaches, as well as RANSAC-, 4PCS-, and BnB-based optimization strategies. The paper benchmarks datasets and metrics, compares representative methods, and discusses challenges such as robustness to outliers, scale variation, and large-scale scene alignment. Collectively, it highlights how learning-based approaches are increasingly integrated with traditional geometric techniques, and it identifies unsupervised and end-to-end paradigms as fertile directions for future research with practical impact in robotics, SLAM, and 3D reconstruction.

Abstract

3D point cloud registration is a fundamental problem in computer vision, computer graphics, robotics, remote sensing, and etc. Over the last thirty years, we have witnessed the amazing advancement in this area with numerous kinds of solutions. Although a handful of relevant surveys have been conducted, their coverage is still limited. In this work, we present a comprehensive survey on 3D point cloud registration, covering a set of sub-areas such as pairwise coarse registration, pairwise fine registration, multi-view registration, cross-scale registration, and multi-instance registration. The datasets, evaluation metrics, method taxonomy, discussions of the merits and demerits, insightful thoughts of future directions are comprehensively presented in this survey. The regularly updated project page of the survey is available at https://github.com/Amyyyy11/3D-Registration-in-30-Years-A-Survey.

Figures (29)

• Figure 1: Typical 3D registration problems.
• Figure 2: A taxonomy of 3D point cloud registration methods.
• Figure 3: Pipeline of correspondence-based 3D pairwise coarse registration.
• Figure 4: Taxonomy of representative geometric methods for 3D pairwise coarse registration.
• Figure 5: Illustration of 3D local descriptor construction pipeline.