Automated Water Finders for Radar Imagery

Abstract

An automated method for findings and extracting water regions from Synthetic Aperture Radar (SAR) imagery is presented. Three automated water finders are implemented for experimentation with two sets of SAR imagery obtained from the areas of Elizabeth City and Asheboro, North Carolina. The input radar image is first segmented into four categories of terrain features. They are water, field, forests, and built-up areas. Water regions from the segmented image are extracted and the remaining image is eliminated. Noise appearing on both water and non-water regions is smoothed and thus completes the automated water finding task. A second water finder with a relatively higher threshold for the region growing process was designed for extracting shallow water regions from some SAR images. Discrimination of non-water features having water like gray values is accomplished by using a third water finder equipped with an additional capability of computing and examining region properties.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 03, 1992
Accession Number
ADA254234

Entities

People

  • Pi-fuay Chen
  • Tho C. Tran

Organizations

  • Army Geospatial Center

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Aerial Photographs
  • Algorithms
  • Classification
  • Computer Vision
  • Computers
  • Discrimination
  • Engineers
  • Errors
  • Image Processing
  • Image Segmentation
  • Images
  • North Carolina
  • Radar
  • Radar Images
  • Segmented
  • Shallow Water
  • Synthetic Aperture Radar

Readers

  • Coastal and Marine Engineering/Sediment Transport/Hydraulic Engineering
  • Computer Vision.