Multi-Sensor Satellite Data Analysis for Urban Land Use Pattern Recognition

Abstract A global phenomenon, urbanization is characterized by the expansion of cities and rapid growth of urban areas. As cities continue to grow, assessing the distribution and trends in land use is increasingly essential to identify, connect, quantify, validate, and predict the negative consequences of urban areas on the environment. In recent decades, the availability of multi-sensor data has revolutionized and emerged as a promising approach for improving efficiency and precision, and each sensor provides distinct capabilities as well as characteristics that make it possible for an in-depth assessment of urban land use. In this context, this research aims to understand the optical data for urban land use mapping using high-resolution remote sensing satellite imagery of different periods. Traditional and machine learning techniques are used to classify urban land use patterns. The results revealed a major land use change with more built-up regions. Urban land area increased from 36.52 % (2012) to 60.23 % (2023), indicating significant urbanization. Additionally, random forest outperformed methods in comparison analysis. Moreover, the significance of this present study is that it offers insightful information into structural and spectral features of urban areas and emphasizes how precise land use classification is crucial for efficient land management, urban planning, and sustainable urban development. The study also exhibits different sources of data from different types of sensors, indicating the possible usage under any condition for emergency monitoring and immediate response.