Utilization of artificial intelligence and thermal cameras in material analysis for hot-summer Mediterranean climates

Abstract This study aims to evaluate the thermal behaviors of surface materials in arid climates to enhance environmental sustainability and energy efficiency. Conducted over 1 year at Dokumapark in Antalya, Turkey, it examines surface temperatures of asphalt, concrete, granite, wood, grass, and soil using thermal using a FLIR-C5 thermal camera. Measurements were taken in the morning, noon, and evening, capturing images from sunny and shaded areas, which were processed with custom Python software. A total of 1728 temperature values were statistically and visually analyzed based on surface–air temperature differences. Seven machine learning models were used for evaluation, with the neural network model achieving the highest accuracy ( R 2 : 0.9848) and minimal error. The model assessed thermal variations across different periods. Grass and wood exhibited low heat retention, while asphalt and brick reached higher temperatures, with asphalt predicted to exceed 50 o C in summer, potentially impacting thermal comfort. Grass was the most efficient material with minimal temperature fluctuations. This study highlights the importance of thermal properties in enhancing energy efficiency and user comfort, as well as the necessity of selecting materials for sustainable cities. It suggests that combining artificial intelligence and thermal imaging techniques can be a beneficial tool for ecological and sustainable architectural design.