Spatial Geometry Analysis of Roadside LiDAR for Improved Vehicle Clustering Accuracy.

Area of Science:

• Intelligent Transportation Systems (ITSs)
• LiDAR technology
• Point Cloud Data Processing

Background:

• Roadside LiDAR is crucial for ITSs, providing high-precision spatial data for traffic monitoring.
• Extracting traffic information from LiDAR point clouds is challenging due to angular sampling, affecting point density with radius and beam distribution.
• Existing methods struggle with varying point densities inherent in LiDAR measurements.

Purpose of the Study:

• To propose a geometry-aware framework for processing roadside LiDAR data.
• To improve the accuracy and stability of traffic information extraction from LiDAR point clouds.
• To address challenges posed by LiDAR's angular sampling and varying point densities.

Main Methods:

• Developed a geometry-aware framework incorporating LiDAR sampling geometry into neighborhood criteria for point-to-point association.
• Defined neighborhood tolerance as a function of radial distance and vertical angular separation.
• Integrated deployment constraints (sensor height, region-of-interest) for physically meaningful connectivity.

Main Results:

• The proposed approach significantly improves clustering accuracy and stability in LiDAR point cloud processing.
• Reduced false negatives in sparse regions and avoided excessive cluster merging in dense areas.
• Demonstrated robust performance across diverse sensing conditions with negligible computational overhead.

Conclusions:

• The geometry-aware framework offers a more accurate and stable method for extracting traffic information from roadside LiDAR data.
• The approach effectively handles the inherent challenges of LiDAR sampling geometry.
• This method provides a significant advancement for intelligent transportation systems utilizing LiDAR sensing.