Revealing the Structure of Land Ownership through the Automatic Vectorisation of Swiss Cadastral Plans

This paper explores how the dynamics and structure of land ownership can be studied by automating the vectorisation of historical cadastral sources. Cadastres are a granular, relatively homogeneous and abundant source. From the 18th century, they are also relatively well aligned with the paradigms of modern geodata. Digital or quantitative methods seem therefore particularly well suited to approach these substantial archival corpora. Our analysis focuses mainly on three cadastres of Lausanne (1722, 1831, 1883), which total 570 plats. The plans are first georeferenced, before being semantically segmented using a neural model, and vectorised. Compared to a manual approach, automation reduces the vectorisation workload by 12 fold. We study the dynamics of persistence by deploying a spatial matching methodology, geometrically detecting parcel fusions and divisions across temporal layers. We also investigate the social structure of ownership in Lausanne at the beginning of the 19th century by relating representative owners to their socioprofessional status based on historical population censuses.