Urban surface classification using self-supervised domain adaptive deep learning models and its application in urban environment studies

Abstract Changed urban surface and human activities in urban areas have led to serious environmental problems globally, including deteriorated local thermal/wind environments and air pollution. In this study, we proposed and validated a domain adaptive land cover classification model, to automatically classify Google Earth images into pixel-based land cover maps. By combining the domain adaptation and self-supervised learning technique, we extend the model’s generalization ability even trained with a small dataset. Furthermore, we explored the impact of land surface composition on the local meteorological parameters and air pollutant concentrations. Urban surface information was extracted around weather stations and air quality monitoring stations from three most developed urban agglomerations in China, including Beijing, Shanghai and the great bay area (GBA). Correlation analysis results show that air temperature has a strong positive correlation with neighbor artificial impervious surface fraction, with Pearson correlation coefficients higher than 0.6 in all areas except for the spring in the GBA. The correlation is much weaker and variant for air pollutants. This work provides an efficient method for urban land cover extraction which will be of great value for assessing the urban surface composition, quantifying the impact of land use/land cover, and facilitating the development of informed policies.