AI-for-Land-Use-Deforestation-Monitoring

The Repository is for an AI model for Land Use & Deforestation Monitoring This Project presents the development of a machine learning model utilizing satellite imagery for near real-time deforestation monitoring in Mabira Forest, Uganda. The model employs a Convolutional Neural Network to perform binary classification and semantic segmentation, distinguishing forested from deforested areas with high accuracy. By leveraging open-source satellite data from platforms like Sentinel 2 and incorporating vegetation indices such as NDVI, the model achieves robust detection of deforestation events. The implementation includes a cloud-based deployment with a user-friendly web interface for stakeholders. The model achieved an accuracy of more than 85%, significantly reducing response times to illegal logging activities. This framework supports Uganda’s National Development Plan and UN Sustainable Development Goals 13 and 15, offering a scalable solution for forest conservation in Uganda and beyond. The Dataset link is here: https://www.kaggle.com/code/mutesasirabashir/ai-for-land-use-deforestation-monitoring/input