A Crowd Movement Analysis Method Based on Radar Particle Flow

Area of Science:

• Public Safety Technology
• Sensor Data Analysis
• Fluid Dynamics Analogy

Background:

• Crowd movement analysis (CMA) is crucial for public safety, aiding in hazard identification.
• Traditional video surveillance struggles with environmental factors like lighting and depth of field.
• Millimeter-wave radar offers all-weather capabilities for accurate distance and angle measurements, overcoming video limitations.

Purpose of the Study:

• To propose a robust crowd motion analysis method using radar data.
• To overcome the limitations of traditional video-based crowd analysis.
• To accurately characterize crowd aggregation, dispersion, and normal movement states.

Main Methods:

• Extraction of radar particle flow (RPF) from adjacent millimeter-wave radar point sets using optical flow.
• Definition of a 'micro-source' concept to analyze RPF vector origins and destinations.
• Calculation of local diffusion potential based on micro-source counts to represent crowd movement states.

Main Results:

• The proposed RPF algorithm effectively analyzes crowd movements in real-world scenarios.
• The method accurately identifies aggregation, dispersion, and normal crowd movements.
• Achieved an accuracy rate of no less than 88% in real-world validation.

Conclusions:

• Radar particle flow provides a reliable method for crowd movement analysis.
• The RPF approach offers superior performance compared to traditional video-based methods, especially in challenging conditions.
• This technology enhances the ability to monitor and manage crowds for improved public safety.