Publication Type

Journal Article

Version

submittedVersion

Publication Date

1-2027

Abstract

Unsupervised LiDAR motion estimation has attracted significant attention in pattern recognition and robotic vision. However, ensuring robust performance under adverse conditions — particularly structured noise like heavy snowfall — remains a challenging problem. Existing models struggle with domain generalization due to the sensitivity of feature extractors to spatial outliers, limiting their reliability in complex real-world scenarios. In this work, we present an unsupervised LiDAR motion estimation framework via patch-aware feature reweighting to mitigate the domain shift between clear and snowy conditions. Our approach prioritizes reliability-aware reweighting to reduce the influence of snowflake noise and outlier points on pose estimation, while balancing computational efficiency for real-time applications. To achieve this, we introduce a Patch Spatial Measure (PSM) module that evaluates point dispersion within patches to identify sparse and scattered noise. Furthermore, we propose a Patch Point Weight Predictor (PPWP) to assign adaptive confidence scores to individual points, enhancing the discriminative capability of local features. In this module, we leverage multi-modal feature fusion to refine point-wise weight prediction and employ an intensity-guided mechanism to rapidly suppress near-field snow with low computational overhead. Our model is trained under clear weather data and rigorously tested across diverse scenarios, including snowy and dynamic environments. Extensive experiments show that our method achieves improved robustness and generalization under snowy conditions without relying on ground-truth supervision.

Keywords

Motion estimation, Patch-aware feature reweighting, Snowy condition, Unsupervised

Discipline

Artificial Intelligence and Robotics | Graphics and Human Computer Interfaces

Research Areas

Intelligent Systems and Optimization

Areas of Excellence

Digital transformation

Publication

Pattern Recognition

Volume

181

First Page

1

Last Page

13

ISSN

0031-3203

Identifier

10.1016/j.patcog.2026.114630

Publisher

Elsevier

Additional URL

https://doi.org/10.1016/j.patcog.2026.114630

Download

Share

COinS