Roof Classification

Roof classification, using various remote sensing data like satellite and drone imagery as well as LiDAR point clouds, aims to automatically identify and categorize roof types for diverse applications. Current research emphasizes the use of deep learning models, including convolutional neural networks (CNNs) and U-Nets, often coupled with techniques like instance segmentation and multi-task learning, to improve accuracy and efficiency. These advancements are crucial for applications ranging from assessing solar energy potential and malaria risk to improving building information for disaster response and urban planning, ultimately contributing to more informed decision-making in various sectors.