SANDRO (Splitting strategy for point cloud Alignment using Non-convex anD Robust Optimization) is a novel algorithm for point cloud registration. It integrates an Iteratively Reweighted Least Squares (IRLS) framework with a Graduated Non-Convexity (GNC) approach and a Geman-McClure robust loss function to handle high outlier rates and skewed outlier distributions.
A key feature of SANDRO is its splitting strategy, which partitions the point cloud into smaller subsets to reduce bias from symmetrical outliers and improve convergence. This technique allows SANDRO to handle complex registration problems that often cause failures in traditional methods.
Unlike traditional methods that struggle with point cloud symmetries and high outlier rates, SANDRO achieves superior accuracy and robustness.
โ The paper has been accepted to ICRA 2025 and it is available here!!
๐ง Installation instructions will be available soon
๐ง Usage examples will be provided soon
If you use SANDRO in your research, please cite:
@article{adlerstein2025sandro,
title={SANDRO: a Robust Solver with a Splitting Strategy for Point Cloud Registration},
author={Michael Adlerstein, Joaฬo Carlos Virgolino Soares, Angelo Bratta, Claudio Semini},
journal={ICRA 2025},
year={2025}
}