Rome organized her Empire via a network of cities, crucial in administration, social organization and economy. By the first century CE there were about 2,000 cities, and understanding them is central to knowledge of the Roman world. However, cities are inherently large objects only very small samples of which can ever be excavated. Study of Roman urbanism has long been dominated by a few towns that have been the subject of major excavations but are hardly typical. Today geophysical survey methods, in particular ground-penetrating radar, can reveal complete, detailed plans of buried towns. However, the precise interpretation and visualization of this growing amount of data remains a major concern to archaeologists who want to contribute to the historical debate on ancient urbanism. This project aims to achieve methodological innovation in 3D- mapping, interpretation and visualization of completely surveyed Roman cities. It will do so by developing computer-aided object detection based on expert knowledge of the revealed buried evidence, and by refining 3D-modelling for the virtual representation of the ancient cityscapes derived from the archaeological evidence. By facilitating mapping, interpretation and visualization of the results from large surveys, the project will stimulate non-invasive investigation and preservation of important cultural heritage sites, help to increase the number of towns with entirely known plan, and change how Roman urbanisation is understood.