Vehicle Detection under Adverse Weather from Roadside LiDAR Data

Area of Science:

• Transportation Engineering
• Sensor Technology
• Environmental Monitoring

Background:

• Roadside light detection and ranging (LiDAR) is a key technology for traffic data collection.
• Existing LiDAR data processing algorithms often fail under adverse weather conditions like wind and snow.
• Reliable traffic data collection is crucial for intelligent transportation systems.

Purpose of the Study:

• To evaluate the performance of current roadside LiDAR data processing algorithms under adverse weather.
• To develop and validate an improved method for processing roadside LiDAR data in challenging weather.
• To enhance the accuracy of vehicle detection using LiDAR in windy and snowy environments.

Main Methods:

• Performance analysis of state-of-the-art roadside LiDAR data processing algorithms.
• Development of a novel background filtering and object clustering technique.
• Experimental testing and validation of the improved method under simulated adverse weather conditions.

Main Results:

• The proposed background filtering and object clustering method significantly improved data processing accuracy.
• The enhanced approach demonstrated superior performance compared to existing state-of-the-art methods in windy and snowy conditions.
• Vehicle identification accuracy was maintained at relatively high levels despite adverse weather.

Conclusions:

• The developed method offers a robust solution for roadside LiDAR data processing in adverse weather.
• This advancement enables more reliable traffic data collection, crucial for intelligent transportation systems.
• The improved accuracy in vehicle detection has significant implications for traffic management and safety.