A Fusion Approach for Water Area Classification Using Visible, near Infrared and Synthetic Aperture Radar for South Asian Conditions

Public Deposited
Analytics

Downloadable Content

Download PDF
Request Version for Screen Reader
Creator
  • Ahmad, S.K.
    • Other Affiliation: University of Washington
  • Hossain, F.
    • Other Affiliation: University of Washington
  • Eldardiry, H.
    • Other Affiliation: University of Washington
  • Pavelsky, T.M.
    • Affiliation: College of Arts and Sciences, Department of Geological Sciences
Abstract
  • Consistent estimation of water surface area from remote sensing remains challenging in regions such as South Asia with vegetation, mountainous topography, and persistent monsoonal cloud cover. High-resolution optical imagery, which is often used for global inundation mapping, is highly impacted by clouds, while synthetic aperture radar (SAR) imagery is not impacted by clouds and is affected by both topographic layover and vegetation. Here, we compare and contrast inundation extent measurements from visible (Landsat-8 and Sentinel-2) and SAR (Sentinel-1) imagery. Each data type (wavelength) has complementary strengths and weaknesses which were gauged separately over selected water bodies in Bangladesh. High-resolution cloud-free PlanetScope imagery at 3-m resolution was used as a reference to check the accuracy of each technique and data type. Next, the optical and radar images were fused for a rule-based water area classification algorithm to derive the optimal decision for the water mask. Results indicate that the fusion approach can improve the overall accuracy by up to 3.8%, 18.2%, and 8.3% during the wet season over using the individual products of Landsat8, Sentinel-1, and Sentinel-2, respectively, at three sites, while providing increased observational frequency. The fusion-derived products resulted in overall accuracy ranging from 85.8% to 98.7% and Kappa coefficient varying from 0.61 to 0.83. The proposed SAR-visible fusion technique has potential for improving satellite-based surface water monitoring and storage changes, especially for smaller water bodies in humid tropical climate of South Asia.
Date of publication
Keyword
DOI
Identifier
Resource type
  • Article
Rights statement
  • In Copyright
Journal title
  • IEEE Transactions on Geoscience and Remote Sensing
Journal volume
  • 58
Journal issue
  • 4
Page start
  • 2471
Page end
  • 2480
Language
  • English
ISSN
  • 0196-2892
Publisher
  • Institute of Electrical and Electronics Engineers Inc.

Relations

Parents:

This work has no parents.

In Collection:

Items

Thumbnail Title Date Uploaded Visibility Actions
file details A_Fusion_Approach_for_Water_Area_Classification_Using_Visible_Near_Infrared_and_Synthetic_Aperture_Radar_for_South_Asian_Conditions.pdf 2022-08-02 Public Download