Historical Map

Historical map analysis focuses on extracting valuable geographic information from scanned maps, often overcoming challenges like inconsistent styles and low-resolution images. Current research employs deep learning models, including U-Net variations and transformers, often incorporating techniques like contrastive learning and semi-supervised learning to improve accuracy and efficiency in tasks such as road network reconstruction, building footprint extraction, and text recognition. These advancements enable large-scale analysis of historical geographic data, providing crucial insights for urban planning, transportation studies, and historical research by automating previously laborious manual processes.