A Fully Interpretable Statistical Approach for Roadside LiDAR Background Subtraction

2025-10-25

AI Summary

Accurately and interpretably separating background from foreground objects in roadside LiDAR point clouds remains challenging. Method: This paper proposes a training-free statistical modeling approach that constructs a Gaussian Distribution Grid (GDG) to uniformly represent the spatial distribution of background points across diverse LiDAR types (e.g., 360° multi-line and MEMS LiDARs), and designs a lightweight filtering algorithm based on statistical distance for point-wise classification. Contribution/Results: The method requires only a small number of background samples, ensures strong interpretability, exhibits cross-sensor generalizability, and imposes minimal hardware requirements. Evaluated on the RCooper dataset, it outperforms existing state-of-the-art methods in accuracy while enabling greater deployment flexibility—making it particularly suitable for large-scale vehicle-infrastructure cooperative perception systems.

Technology Category

• Intelligent Robots: State Estimation
• Machine Learning: Calibration & Uncertainty Quantification
• Planning, Routing, and Scheduling: Optimization of Spatio-temporal Systems

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

We present a fully interpretable and flexible statistical method for background subtraction in roadside LiDAR data, aimed at enhancing infrastructure-based perception in automated driving. Our approach introduces both a Gaussian distribution grid (GDG), which models the spatial statistics of the background using background-only scans, and a filtering algorithm that uses this representation to classify LiDAR points as foreground or background. The method supports diverse LiDAR types, including multiline 360 degree and micro-electro-mechanical systems (MEMS) sensors, and adapts to various configurations. Evaluated on the publicly available RCooper dataset, it outperforms state-of-the-art techniques in accuracy and flexibility, even with minimal background data. Its efficient implementation ensures reliable performance on low-resource hardware, enabling scalable real-world deployment.

Problem

• Developing interpretable background subtraction for roadside LiDAR data
• Classifying LiDAR points as foreground or background statistically
• Supporting diverse LiDAR types and low-resource hardware deployment

Innovation

• Uses Gaussian distribution grid for background modeling
• Filters LiDAR points via statistical classification algorithm
• Supports diverse LiDAR types and low-resource hardware