A knowledge-based approach to raster-vector conversion of large scale topographic maps
AbstractPaper-based raster maps are primarily for human consumption, and their interpretation always requires some level of human expertese. Todays computer services in geoinformatics usually require vectorized topographic maps. The usual method of the conversion has been an error-prone, manual process. In this article, the possibilities, methods and difficulties of the conversion are discussed. The results described here are partially implemented in the IRIS project, but further work remains. This emphasizes the tools of digital image processing and knowledge-based approach. The system in development separates the recognition of point-like, line-like and surface-like objects, and the most successful approach appears to be the recognition of these objects in a reversed order with respect to their printing. During the recongition of surfaces, homogeneous and textured surfaces must be distinguished. The most diverse and complicated group constitute the line-like objects. The IRIS project realises a moderate, but significant step towards the automatization of map recognition process, bearing in mind that full automatization is unlikely. It is reasonable to assume that human experts will always be required for high quality interpretation, but it is an exciting challenge to decrease the burden of manual work.
Download data is not yet available.
How to Cite
Szendrei, R., Elek, I., & Márton, M. (2011). A knowledge-based approach to raster-vector conversion of large scale topographic maps. Acta Cybernetica, 20(1), 145-165. https://doi.org/10.14232/actacyb.20.1.2011.11