A Survey on Satellite Image Processing Using Machine Learning

This survey examines the application of deep learning to satellite imagery for advanced Earth observation analysis, focusing on land use, object detection, and land mapping. The project employs Convolutional Neural Networks (CNNs) to achieve accurate LULC classification, object detection (e.g., swimming pools, cars), and image segmentation tasks. Key applications include multi-class land cover mapping, binary segmentation for building identification, and landslide detection, all of which play an essential role in accurately defining land features. By automating these analyses, the project facilitates rapid, scalable assessments that support timely and informed decision-making in sectors like agriculture, forestry, and disaster management, ultimately enhancing geographic understanding and resource management.