Classification of urban functional areas from remote sensing images and time-series user behaviour data

Urbanization is accelerating at a rapid rate, which has introduced many challenges, especially in the field of urban planning. Under the backdrop of global urbanization, some cities are particularly vulnerable to climate change and natural disasters that are influenced by unplanned urban expansion. Rational planning of urban functional areas needs to be strengthened to improve the scientific approach of urban planning and urbanization. In this study, the classification of urban functional areas based on dual-modal data (i.e., remote sensing image and user behavior data) was implemented using machine learning (ML) algorithms. After the set test, the classification accuracy of urban functional areas reached 82.45%. Through analysis, it could be concluded that the use of data of two modalities achieved a higher classification accuracy than that achieved by using data of a single modality. The data of the two modalities complement each other, and the use of ML algorithms to train such data can yield good results.