Comparing the urban environment with socioeconomic characteristics using features extracted from aerial imagery
Melanie Green and Dani Arribas-Bel
Information about the characteristics and activities of humans becomes encoded in the landscape as it is continually shaped by the people who live there. These socioeconomic and demographic characteristics are traditionally measured using census data and represented using geodemographic classifications. Remotely sensed images at a high resolution contain a huge amount of detail on the urban environment and its potential to reveal complex patterns is unlocked by deep learning. This paper uses convolutional neural networks to extract features from aerial imagery and uses clustering to classify areas and compare their physical features to their socioeconomic features.