Accession Number : ADA311189

Title :   Automatic Extraction of Drainage Network from Digital Terrain Elevation Data (DTED).

Descriptive Note : Final rept.,

Corporate Author : MAINE UNIV AT ORONO

Personal Author(s) : Musavi, Mohamad T. ; Hawkins, Greg ; Fern, Alan ; Ruffing, Curt

PDF Url : ADA311189

Report Date : MAY 1996

Pagination or Media Count : 33

Abstract : This report describes an approach to the automatic extraction of a drainage network from Digital Terrain Elevation Data (DTED). The extraction of drainage networks from topological data is one of the more important uses of such information. An extracted network may be used in many applications, such as determining drainage network metrics or selecting control points for image registration and mapping applications. Neural networks have been shown to possess generalization properties that make them well suited to this problem, which calls for a generalized solution. In this research, a biologically inspired network has been designed and applied to automatic extraction of drainage networks. Several DTED test files were used for testing of the developed model. The results indicate the success of neural networks in the extraction of drainage networks from DTED.

Descriptors :   *NEURAL NETS, *DRAINAGE, *DIGITAL MAPS, DATA BASES, TERRAIN, TOPOGRAPHY, TOPOLOGY, ELEVATION, AUTOMATIC, IMAGE REGISTRATION.

Subject Categories : Cartography and Aerial Photography

Distribution Statement : APPROVED FOR PUBLIC RELEASE