Abstract
Semi-automated DEM reconstruction based on historical maps
Track: Mapping and Visualization
Authors: Samuel Wiesmann
Historical printed maps may store spatial information of the past, e.g. the shape of the landscape by means of contour lines. Digitizing this information manually is time consuming, still more if contour lines must be attributed with height values.
Our fully ArcGIS-based approach facilitates this process, starting with a color image segmentation of the scanned image in ArcMap. Using ArcScan, the separated colors are then vectorized. To accelerate attribution of derived contour lines, an ArcPy-based Tool was developed, which determines the correct order of selected line geometries and writes calculated height values into the attribute table. Contour lines are then used to interpolate the DEM, which enables further analysis and visualization.
The color image segmentation can easily be transferred to other images. The attribution tool is applicable not only for height attribution, but also generally for sorting objects based on their location contrary to sorting based on a known attribute.