Fractal Terrain Analysis from an Attractor Viewpoint
Abstract
We investigated terrain modeling based upon fractal technique and scale space fingerprints. Terrain can be described by features (such as ridges, valleys and slopes) that are fundamentally edges in topography. Since scale space fingerprints extract edges from a terrain profile in an order of importance based on edge strength, they provide a natural approach for terrain analysis. Therefore, we applied scale space fingerprints as a primary analysis tool to Digital Elevation Models of various terrain types. Scale space fingerprint analysis yields two important results for terrain modeling: First, by using a feature density function that describes the number of terrain edges as a function of smoothing scale, we were able to discriminate between different landforms and even between different expressions of the same landform. And, using the same feature density function as a descriptor, we were able to determine fractal behavior of a landform more sensitively than by conventional methods, and without a dependence on measuring units that can mask fractal behavior. We also initiated theory development for a two dimensional fingerprint capability based on the gaussian curvature operator. (jg)
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 26, 1990
- Accession Number
- ADA219620
Entities
People
- M. A. Piech
Organizations
- State University of New York at Albany