Towards Explicit Geometry-Reflectance Collaboration for Generalized LiDAR Segmentation in Adverse Weather

2025-06-03

AI Summary

LiDAR semantic segmentation accuracy degrades significantly under adverse weather conditions due to domain shifts in point cloud geometry and intensity. Method: This paper identifies a synergistic structural–intensity shift mechanism across heterogeneous weather domains and proposes the first geometry–intensity dual-branch decoupled network to independently model their distinct degradation patterns. It introduces multi-scale feature alignment and uncertainty-aware fusion modules to enable robust, collaborative feature suppression, all within an end-to-end learning framework that requires no weather simulation or data augmentation. Contribution/Results: The method achieves new state-of-the-art performance on multiple adverse-weather LiDAR benchmarks, with substantially improved cross-weather generalization—yielding a +4.2% mIoU gain over prior best approaches.

Technology Category

• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Computer Vision: Adversarial Attacks & Robustness
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Existing LiDAR semantic segmentation models often suffer from decreased accuracy when exposed to adverse weather conditions. Recent methods addressing this issue focus on enhancing training data through weather simulation or universal augmentation techniques. However, few works have studied the negative impacts caused by the heterogeneous domain shifts in the geometric structure and reflectance intensity of point clouds. In this paper, we delve into this challenge and address it with a novel Geometry-Reflectance Collaboration (GRC) framework that explicitly separates feature extraction for geometry and reflectance. Specifically, GRC employs a dual-branch architecture designed to independently process geometric and reflectance features initially, thereby capitalizing on their distinct characteristic. Then, GRC adopts a robust multi-level feature collaboration module to suppress redundant and unreliable information from both branches. Consequently, without complex simulation or augmentation, our method effectively extracts intrinsic information about the scene while suppressing interference, thus achieving better robustness and generalization in adverse weather conditions. We demonstrate the effectiveness of GRC through comprehensive experiments on challenging benchmarks, showing that our method outperforms previous approaches and establishes new state-of-the-art results.

Problem

• Addresses LiDAR segmentation accuracy drop in adverse weather
• Explores domain shifts in point cloud geometry and reflectance
• Proposes dual-branch feature extraction for robust segmentation

Innovation

• Dual-branch architecture for geometry and reflectance
• Multi-level feature collaboration module
• Explicit separation of geometric and reflectance features