DMNR: Unsupervised De-noising of Point Clouds Corrupted by Airborne Particles

TL;DR

This paper introduces two unsupervised filtering methods, DMNR and DMNR-H, that effectively remove airborne particle noise from LiDAR point clouds, improving perception in adverse weather conditions without requiring laborious annotations.

Contribution

The paper presents novel dynamic filtering algorithms, DMNR and DMNR-H, that outperform existing unsupervised methods and rival supervised approaches in denoising LiDAR data affected by airborne particles.

Findings

DMNR and DMNR-H outperform state-of-the-art unsupervised methods.

They are slightly better than supervised deep learning methods.

The methods are robust across different LiDAR sensors and environmental conditions.

Abstract

LiDAR sensors are critical for autonomous driving and robotics applications due to their ability to provide accurate range measurements and their robustness to lighting conditions. However, airborne particles, such as fog, rain, snow, and dust, will degrade its performance and it is inevitable to encounter these inclement environmental conditions outdoors. It would be a straightforward approach to remove them by supervised semantic segmentation. But annotating these particles point wisely is too laborious. To address this problem and enhance the perception under inclement conditions, we develop two dynamic filtering methods called Dynamic Multi-threshold Noise Removal (DMNR) and DMNR-H by accurate analysis of the position distribution and intensity characteristics of noisy points and clean points on publicly available WADS and DENSE datasets. Both DMNR and DMNR-H outperform state-of-the-art unsupervised methods by a significant margin on the two datasets and are slightly better than supervised deep learning-based methods. Furthermore, our methods are more robust to different LiDAR sensors and airborne particles, such as snow and fog.