Automatic Vectorization of Historical Maps

Abstract

For the most diverse practical purposes and communication needs (from trade to conquests or exploratory voyages), man has always represented space through the production of maps. There are countless varieties of themes, styles and graphic languages, types of representation (pictorial maps, bird's eye views, cadastral and topographic maps) used throughout history to depict, in particular, the landscape and the urban fabric. The informative, historical, cultural and identity value of this cartographic heritage, now preserved in archives and institutions around the world, is inestimable. Due to the need to preserve its content from the ravages of time and to allow its "digital" use, an increasingly intense campaign of digitization of these sources has been launched in recent years (as has happened, for example, for other historical data, such as photographs and books). The growing availability of digital historical maps and, at the same time, the notable progress in data processing and analysis, today offers new perspectives for the use and valorization of these resources. In the case, in particular, of ancient cartography, there are increasingly broader scenarios for accurate studies on the evolution and transformation of urban structures, to support careful urban planning policies or reconstruct historical scenarios (Farella et al., 2021).

Keywords

• Cartography
• Historical Maps
• GIS
• Automatic vectorization
• artificial intelligence
• geometric data