Cartographic Data Entry through Automatic Feature Tracking

Many cartographic systems currently rely on raster-scan digitizing to convert analog source material to digital form. Raster technology is very effective at rapid and accurate digitization of large volumes of cartographic data. At the same time, existing raster-to-vector (R-V) conversion processes rely on an inordinately large amount of human post-scan editing to coherently sort and combine the short, unattributed lineal segments into single cartographic spatial entities. The Automatic Feature Tracking (AFT) system addresses one of the major causes of this bottleneck in the digitization process—skeletonization. This is accomplished by directly converting symbolized linear features on a raster map image into sets of x,y coordinates. The system relies on Template Matching and Feature Tracking techniques to locate feature centerlines. This paper briefly reviews the history of map digitization techniques, illustrates inadequacies of those past approaches, and presents the AFT system as an alternative to the R-V conversion routines in existing raster digitization systems.