Augmenting grid-based contours to improve thin plate dem generation. Photogrammetric Engine.pdf

Augmenting Grid-based Contours to Improve Thin Plate DEM Generation April 4, 2003 Michael B. Gousie Department of Mathematics Computer Science Wheaton College Norton, MA 02766 Wm. Randolph Franklin Electrical, Computer, and Systems Engineering Department Rensselaer Polytechnic Institute Troy, NY 12180 Abstract We present two new pre-processing techniques that improve thin plate Digital El- evation Model (DEM) approximations from grid-based contour data. One method computes gradients from an initial interpolated or approximated surface. The aspects are used to create gradient lines that are interpolated using Catmull-Rom splines. The computed elevations are added to the initial contour data set. Thin plate methods are applied to all of the data. The splines allow information to flow across contours, improving the final surface. The second method successively computes new, interme- diate contours in between existing isolines, which provide additional data for subse- quent thin plate processing. Both methods alleviate artifacts visible in previous thin plate methods. The surfaces are tested with published methods to show qualitative and quantitative improvements over previous methods. 2 1 Introduction Geographical Information Systems (GIS) are becoming increasingly popular for visual- izing spatial data. Most systems layer patterns or colors, which depict data such as soil type, roads, and the like, over a two-dimensional map. As technology continues to im- prove, users increasingly expect to view such data in three-dimensions, as is now done in ArcView and MapInfo. The user can then view the desired data in the context of the surrounding topology. Digital Elevation Models (DEM) are often used to store three-dimensional elevation data via a regular grid. Because DEMs are not available for many areas and/or because they are storage intensive, they are often interpolated or approximated from sparse data. We have chosen isoline data from which to compute DEMs because con