• DocumentCode
    814140
  • Title

    DNESYS-an expert system for automatic extraction of drainage networks from digital elevation data

  • Author

    Qian, Jianzhong ; Ehrich, Roger W. ; Campbell, James B.

  • Author_Institution
    Virginia Polytech. Inst. & State Univ., Blacksburg, VA, USA
  • Volume
    28
  • Issue
    1
  • fYear
    1990
  • fDate
    1/1/1990 12:00:00 AM
  • Firstpage
    29
  • Lastpage
    45
  • Abstract
    The problems that arose in earlier programs to map drainage systems are analyzed in detail. An expert system called the drainage Network Extraction System (DNESYS) is described. It uses both local operator and global reasoning to extract drainage networks and ridge lines. A stream representation called a parameterized directed graph (PDG) is constructed to model a drainage system. The construction of the model begins with an initial pixel labeling procedure. Then, a network tracing and property measurement procedure converts the two-dimensional low-level labeling information into a symbolic database for high-level processing. The application of Dempster-Shafer evidence theory allows evidence collection and uncertain reasoning to be performed against the DNESYS knowledge-base that contains the drainage system model and the organized expert knowledge. By discarding erroneous information and supplying missing information, DNESYS produces a complete PDG which can be converted into the final drainage system
  • Keywords
    cartography; computerised picture processing; database management systems; geophysics computing; knowledge based systems; rivers; topography (Earth); DNESYS; Dempster-Shafer evidence theory; automatic extraction; cartography; digital elevation data; drainage Network Extraction System; drainage systems; erroneous information; evidence collection; expert system; global reasoning; high-level processing; knowledge-base; local operator; mapping; missing information; network tracing; parameterized directed graph; pixel labeling procedure; property measurement procedure; ridge lines; river; stream representation; symbolic database; topography; two-dimensional low-level labeling information; uncertain reasoning; Data mining; Databases; Expert systems; Geographic Information Systems; Geologic measurements; Geoscience; Graphics; Labeling; Senior members; Spatial resolution;
  • fLanguage
    English
  • Journal_Title
    Geoscience and Remote Sensing, IEEE Transactions on
  • Publisher
    ieee
  • ISSN
    0196-2892
  • Type

    jour

  • DOI
    10.1109/36.45743
  • Filename
    45743