Predicting Urban Air Temperature Using UAV Infrared Thermography and an Artificial Neural Network

초록

This study introduces a new approach for predicting urban air temperatures using UAV infrared thermography and an artificial neural network (ANN), highlighting its potential application in urban microclimate modeling. Meteorological data and UAV-based surface temperature measurements from diverse urban environments-including asphalt, grass, tree shade, and water bodies-were used for ANN model development. The results show that air temperature prediction accuracy improved when surface temperature was included as an input variable. These findings suggest that incorporating surface temperature can improve the efficiency of predicting air temperatures across diverse urban spaces. © 2025 Building Simulation Conference Proceedings. All rights reserved.