Kriging-based spatial interpolation from measurements for sound level mapping in urban areas

Area of Science:

• Environmental acoustics
• Geographic Information Systems (GIS)
• Spatial statistics

Background:

• Network-based sound monitoring and participatory sensing mobile applications are widespread.
• Sparse spatial and temporal data from these systems challenge the creation of accurate sound maps.

Purpose of the Study:

• To evaluate different spatial interpolation strategies for generating sound maps.
• To investigate the impact of observation density on sound map accuracy.

Main Methods:

• Conducted a mobile measurement campaign in Paris, collecting data by walking every street multiple times.
• Assessed four interpolation strategies: Ordinary Kriging, Universal Kriging, and two network-topology-aware distance metrics.
• Adaptively constructed noise maps based on collected measurements.

Main Results:

• A high density of observation points is essential for producing interpolated sound maps that closely resemble a reference map.
• The performance of interpolation strategies is directly influenced by the density of collected acoustic data.

Conclusions:

• Achieving accurate sound maps requires a substantial density of acoustic measurements.
• Future participatory sensing initiatives must prioritize data density to overcome limitations in spatial interpolation.