Sight View Constraint for Robust Point Cloud Registration

TL;DR

This paper introduces the Sight View Constraint (SVC), a novel method to improve the robustness of partial point cloud registration, especially under low overlap conditions, by effectively identifying incorrect transformations.

Contribution

The paper proposes the Sight View Constraint (SVC), a new general approach that enhances partial point cloud registration robustness by conclusively rejecting incorrect transformations.

Findings

SVC improves registration recall from 78% to 82% on 3DLoMatch.

SVC achieves state-of-the-art results in partial PCR.

Experiments validate SVC's effectiveness on indoor and outdoor scenes.

Abstract

Partial to Partial Point Cloud Registration (partial PCR) remains a challenging task, particularly when dealing with a low overlap rate. In comparison to the full-to-full registration task, we find that the objective of partial PCR is still not well-defined, indicating no metric can reliably identify the true transformation. We identify this as the most fundamental challenge in partial PCR tasks. In this paper, instead of directly seeking the optimal transformation, we propose a novel and general Sight View Constraint (SVC) to conclusively identify incorrect transformations, thereby enhancing the robustness of existing PCR methods. Extensive experiments validate the effectiveness of SVC on both indoor and outdoor scenes. On the challenging 3DLoMatch dataset, our approach increases the registration recall from 78\% to 82\%, achieving the state-of-the-art result. This research also highlights the significance of the decision version problem of partial PCR, which has the potential to provide novel insights into the partial PCR problem.