A Hybrid Approach to Predicting Land Use/Land Cover Change: Integrating Object-Based Detection with Machine Learning

Urbanization in rapidly growing cities like Jaipur has led to significant land use and land cover (LULC) transformations, impacting natural resources and urban ecosystems. Predicting these changes accurately is critical for sustainable urban planning. This study leverages a hybrid approach combining Cellular Automata (CA) and Support Vector Machine (SVM) to model and predict LULC changes in Jaipur over the periods of 2017, 2024, and projected for 2031. Utilizing Sentinel-2 satellite data, spatial variables such as proximity to roads, central business districts, drainage, and slope were integrated to assess their influence on urban growth patterns. The results reveal that built-up areas have expanded considerably, largely replacing open and green spaces, driven by infrastructure proximity and low-slope zones. The CA-SVM model demonstrated strong predictive accuracy, with key findings suggesting continued urban sprawl that could strain local resources and reduce ecological stability by 2031. These insights highlight the value of data-driven planning strategies that can manage urban growth sustainably. Future research could enhance the model by incorporating additional socio-economic variables and employing high-resolution data for improved accuracy. The hybrid CA-SVM approach offers a robust framework for understanding and predicting LULC dynamics, providing a valuable tool for urban policymakers aiming to balance development with environmental conservation.Cite this article as: R. L. Chouhan and H. S. Shekhawat, “A hybrid approach to predicting land use/land cover change: Integrating object-based detection with machine learning,” Electrica, 25, 0181, 2025. doi: 10.5152/electrica.2025.24181.