Urban Sound Propagation

This dataset is assembled for research into urban sound propagation, comprising 25,000 data points across 10 diverse cities. Each city is represented by 2,500 locations, offering a comprehensive analysis of various urban configurations. The dataset utilizes OpenStreetMap (OSM) imagery to detail the urban layout within a 500m x 500m area for each location, where buildings are delineated with black pixels and open spaces with white pixels.

Supplementing the urban structural images, the dataset includes sound distribution maps at resolutions of 512x512 and 256x256. These maps are precisely generated through the NoiseModelling v4.0 framework, an advanced simulation tool engineered for accurate modeling of sound dynamics within urban environments.

The data collection methodology involves sourcing information via the Overpass API, subsequently refined and verified using GeoPandas to ensure the dataset's integrity and consistency. For researchers and experts interested in exploring the intricacies of sound simulation, additional insights can be obtained from the NoiseModelling framework documentation.