Automated Image Matching: An Efficient Tool for Georeferencing Historical Cadastral Maps

The cadastral maps from the first systematic geodetic survey of Bosnia and Herzegovina in the late 19th century have always been highly significant. This survey utilized the graphical method of plane table mapping, involving on-site map preparation through direct field observations. The information on these maps is in high demand among technical and legal experts, and they are increasingly serving as a valuable source for historical research. The unique data contained in these maps plays a crucial role in resolving property boundary disputes, addressing land ownership issues, understanding landscape development throughout history, and conducting various territorial analyses. To thoroughly examine, interpret, and make effective use of old cadastral maps, it is essential to digitize them and precisely georeference their spatial information. In this paper, we assessed a method that employs automated image matching for geometric transformations of scanned cadastral maps. The assessment involved registration of the old cadastral map, originally created at a scale of 1:6250, to more contemporary maps drafted at a scale of 1:5000. The main purpose of the research is to provide a solution for transforming the content of an old map into the modern reference coordinate system through an automated procedure. It was demonstrated that the image-matching procedure has significant potential in terms of swiftly identifying common points in old and modern maps. The resulting set of usable homologous points is suitable for the initial georeferencing of old cadastral maps. Complete automatic detection of all corresponding points is not possible with the used technology due to the distorted and incomplete geometric content of the old maps.