Integrated Framework for Accurate Change Detection and Deforestation Identification in Satellite Imagery

The escalation of deforestation, particularly in the urban periphery, underscores the urgency for advanced frameworks in change detection. This research addresses this imperative by introducing a comprehensive solution for precise identification of deforestation and change detection in satellite imagery. Our proposed approach encompasses two key frameworks: one focused on change detection and the other on deforestation identification. A Modified U-Net for change detection, emphasizing pixel-wise alterations in satellite imagery. Simultaneously, a VGG Net model is employed to identify deforested areas, focusing on distinguishing relevant features. The Modified U-Net architecture leads with an exceptional 99.25% accuracy, excelling in detecting nuanced alterations, especially in urban landscapes. Integration of the VGG Net model further amplifies accuracy, achieving a commendable 91.87% in pinpointing deforested areas. This dual proficiency showcases our framework's adaptability, adeptly handling diverse environmental monitoring challenges. Emphasizing robustness and effectiveness, our framework emerges as a pivotal tool in remote sensing, making significant contributions to environmental conservation and land-use monitoring.