Abstract:Estimating the rigid transformation with 6 degrees of freedom based on a putative 3D correspondence set is a crucial procedure in point cloud registration. Existing correspondence identification methods usually lead to large outlier ratios ($>$ 95 $\%$ is common), underscoring the significance of robust registration methods. Many researchers turn to parameter search-based strategies (e.g., Branch-and-Bround) for robust registration. Although related methods show high robustness, their efficiency is limited to the high-dimensional search space. This paper proposes a heuristics-guided parameter search strategy to accelerate the search while maintaining high robustness. We first sample some correspondences (i.e., heuristics) and then just need to sequentially search the feasible regions that make each sample an inlier. Our strategy largely reduces the search space and can guarantee accuracy with only a few inlier samples, therefore enjoying an excellent trade-off between efficiency and robustness. Since directly parameterizing the 6-dimensional nonlinear feasible region for efficient search is intractable, we construct a three-stage decomposition pipeline to reparameterize the feasible region, resulting in three lower-dimensional sub-problems that are easily solvable via our strategy. Besides reducing the searching dimension, our decomposition enables the leverage of 1-dimensional interval stabbing at all three stages for searching acceleration. Moreover, we propose a valid sampling strategy to guarantee our sampling effectiveness, and a compatibility verification setup to further accelerate our search. Extensive experiments on both simulated and real-world datasets demonstrate that our approach exhibits comparable robustness with state-of-the-art methods while achieving a significant efficiency boost.

Robust Point Set Registration Based on Semantic Information

A New Method for Registration of 3d Point Sets with Low Overlapping Ratios

An Automatic Registration Approach to Laser Point Sets Based on Multidiscriminant Parameter Extraction

A Robust Registration Algorithm Based on Salient Object Detection

Robust Point Set Registration Method Based on Global Structure and Local Constraints

Robust point cloud registration based on both hard and soft assignments

Semantics and Spatial Structures Guided Point Cloud Registration

Robust Rigid Registration Algorithm Based on Pointwise Correspondence and Correntropy

Robust Rgb-D Data Registration Based On Correntropy And Bi-Directional Distance

Fast Semantic-Assisted Outlier Removal for Large-scale Point Cloud Registration

Robust Rigid Registration Algorithm Based on Correntropy and Bi-Directional Distance.

Registration-Is-Evaluation: Robust Point Set Matching With Multigranular Prior Assessment

Precise Point Set Registration with Color Assisted and Correntropy for 3D Reconstruction

Robust Point Cloud Registration Framework Based on Deep Graph Matching

Efficient and Robust Point Cloud Registration via Heuristics-guided Parameter Search

An Effective Approach for Point Clouds Registration Based on the Hard and Soft Assignments

An Automatic and Robust Point Cloud Registration Framework Based on View-Invariant Local Feature Descriptors and Transformation Consistency Verification

Graph Correspondence-Based Point Set Registration

Robust Non-rigid Registration Based on Affine ICP Algorithm and Part-Based Method

Robust Non-Rigid Point Set Registration Based on Dynamic Tree

Quantity-Aware Coarse-to-Fine Correspondence for Image-to-Point Cloud Registration