Improvement and Validation of a Smart Road Traffic Noise Model Based on Vehicles Tracking Using Image Recognition:...

Area of Science:

• Environmental Science
• Acoustics
• Transportation Engineering

Background:

• Road traffic noise is a significant environmental pollutant, particularly in European urban areas.
• Accurate assessment of traffic noise requires reliable input data for simulation models.
• Existing methods for data collection can be labor-intensive or lack precision.

Purpose of the Study:

• To develop and validate a methodology for collecting essential road traffic noise model inputs (vehicle count, category, speed) from video recordings.
• To integrate this data with the CNOSSOS-EU model for accurate noise impact assessment.
• To provide a more efficient and precise method for traffic noise analysis.

Main Methods:

• A Python routine using image inference was developed for instantaneous vehicle detection, speed, and categorization (light/heavy) from video data.
• Vehicle data was collected from an Italian highway video recording.
• The collected data was processed using the CNOSSOS-EU model to estimate noise power levels and overall traffic noise impact.

Main Results:

• The developed methodology demonstrated good performance in collecting traffic input data.
• The noise impact estimation showed a mean error of -1.0 dBA.
• A mean absolute error (MAE) of 3.6 dBA was achieved, indicating high accuracy.

Conclusions:

• The proposed methodology offers an effective approach for gathering crucial data for road traffic noise modeling.
• This method enhances the accuracy and efficiency of noise impact assessments, particularly in urban environments.
• The findings contribute to better management and mitigation strategies for traffic noise pollution.