UQ-Loc: Uncertainty-Aware LiDAR Scene Coordinate Regression

2026-08-06

AI Summary

This work addresses the limitation of existing LiDAR-based scene coordinate regression methods, which produce only deterministic predictions and neglect observation uncertainty, thereby compromising localization robustness. We propose the first approach to model full anisotropic Gaussian covariance for this task, predicting a positive-definite 3×3 covariance matrix per voxel. The model is trained using a negative log-likelihood loss augmented with a kNN-based spatial smoothing regularizer. During inference, we introduce uncertainty-weighted seed scoring and Mahalanobis-distance-based inlier verification to enhance pose estimation accuracy. Furthermore, we devise an uncertainty evaluation protocol based on Expected Calibration Error (ECE) and integrate it into an end-to-end localization pipeline. Experiments demonstrate consistent performance gains in 6-degree-of-freedom localization while yielding well-calibrated uncertainty estimates.

Abstract

LiDAR-based Scene Coordinate Regression (SCR) maps point clouds directly to 3D scene coordinates, enabling precise 6-DoF localisation without explicit map retrieval. However, existing methods produce deterministic predictions, discarding aleatoric uncertainty that could improve robustness and downstream decision-making. We present UQ-Loc, which extends the LightLoc architecture with an anisotropic Gaussian covariance head that predicts a full 3x3 positive-definite covariance matrix per voxel. Training uses a Negative Log-Likelihood (NLL) loss augmented with a kNN-based spatial smoothness regulariser, while inference employs a modified SC2-PCR solver with uncertainty-weighted seed scoring and a Mahalanobis-distance inlier test. We adopt Expected Calibration Error (ECE) as a principled metric for evaluating the quality of the predicted uncertainty. Experiments demonstrate that UQ-Loc achieves consistent improvement in 6-DoF localization accuracy while producing well-calibrated covariances.

Problem

• LiDAR
• Scene Coordinate Regression
• Uncertainty Quantification
• 6-DoF Localization
• Aleatoric Uncertainty

Innovation

• Uncertainty Quantification
• LiDAR Localization
• Scene Coordinate Regression
• Anisotropic Covariance
• Expected Calibration Error