Journal article
An Inductive Knowledge-based Approach to Terrain Feature Extraction
Cartography and Geographic Information Science, Vol.23(1), pp.3-19
1996
DOI: 10.1559/152304096782512177
Abstract
Accurate digital representations of basin morphology are needed to support hydrologic flow and sediment transport models. The derivation of such representations from United States Geodetic Survey (USGS) digital elevation models (DEMs) has been an active area of research. The error inherent in USGS DEM data sets can limit their utility for watershed modeling, however. This article describes a new approach for extracting basin morphology from a DEM. The new approach uses inductive procedures and domain-specific knowledge to identify and correct errors. A bit-mapped classification scheme is used to place each point in a DEM into one of six hydrologic categories based on the topographic form of its local and extended neighborhood. Points with similar topographic form are aggregated into networks that represent stream channels and basin divides. These networks are extracted from the DEM and stored in vector-based data structures. Iterative examination of the collected information generates an increasingly accurate representation of basin morphology.
Details
- Title: Subtitle
- An Inductive Knowledge-based Approach to Terrain Feature Extraction
- Creators
- David A. Bennett - University of IowaMarc P. Armstrong - University of Iowa
- Resource Type
- Journal article
- Publication Details
- Cartography and Geographic Information Science, Vol.23(1), pp.3-19
- DOI
- 10.1559/152304096782512177
- ISSN
- 1523-0406
- Number of pages
- 17
- Language
- English
- Date published
- 1996
- Academic Unit
- Geographical and Sustainability Sciences
- Record Identifier
- 9983557275002771
Metrics
24 Record Views