3D Roadway Scene Object Detection with LIDARs in Snowfall Conditions
Ghazal Farhani, Taufiq Rahman, Syed Mostaquim Ali, Andrew Liu, Mohamed Zaki, Dominique Charlebois, Benoit Anctil
TL;DR
The paper addresses LiDAR perception degradation in snowfall by developing a physics-based attenuation model that quantifies signal loss and near-source backscatter, enabling conversion of clear-weather point clouds into snowy analogs via a defined extinction framework. It combines Rayleigh and Mie-based attenuation theories with a snow-specific size distribution to compute the extinction coefficient $\alpha$ and uses a ratio $\frac{P_{snow}(R)}{P_{clear}(R)}$ to synthesize snowy scenes from clear data. The approach is validated with real snowfall data and a state-of-the-art detector (SphereFormer) to evaluate how snowfall impacts 3D object detection, revealing significant near-field false positives and degraded performance that motivate domain-specific retraining and calibration. The work contributes a physics-informed data synthesis method for adverse-weather LiDAR, supporting more robust perception in autonomous driving, and outlines clear paths for improved calibration and detector adaptation.
Abstract
Because 3D structure of a roadway environment can be characterized directly by a Light Detection and Ranging (LiDAR) sensors, they can be used to obtain exceptional situational awareness for assitive and autonomous driving systems. Although LiDARs demonstrate good performance in clean and clear weather conditions, their performance significantly deteriorates in adverse weather conditions such as those involving atmospheric precipitation. This may render perception capabilities of autonomous systems that use LiDAR data in learning based models to perform object detection and ranging ineffective. While efforts have been made to enhance the accuracy of these models, the extent of signal degradation under various weather conditions remains largely not quantified. In this study, we focus on the performance of an automotive grade LiDAR in snowy conditions in order to develop a physics-based model that examines failure modes of a LiDAR sensor. Specifically, we investigated how the LiDAR signal attenuates with different snowfall rates and how snow particles near the source serve as small but efficient reflectors. Utilizing our model, we transform data from clear conditions to simulate snowy scenarios, enabling a comparison of our synthetic data with actual snowy conditions. Furthermore, we employ this synthetic data, representative of different snowfall rates, to explore the impact on a pre-trained object detection model, assessing its performance under varying levels of snowfall