Coastal Bathymetry from Hyperspectral Data.

Abstract

A study is underway to investigate relationships of water depth, bottom type, and other optical variables to upwelling spectral radiance of coastal waters. A neural network and data from the Airborne Visible/Infrared Imaging Spectrometer (AVIRIS) are used to quantify these relationships. Data is analyzed for two areas: one on the western coast of Florida in Tampa Bay and the other on the Florida panhandle in Santa Rosa Sound. AVIRIS data from Tampa Bay is atmospherically corrected, whereas the data from the Santa Rosa Sound is not atmospherically corrected. The neural network can compute reasonable depths from spectral radiance in both cases. Sounding data obtained from the National Ocean Survey (NOS) hydrographic database is used in the training phase of the neural network and to test the accuracy of the result. Depths estimated by the neural network for Tampa Bay are accurate to a RMS error of 3.9 ft and for the Santa Rosa Sound to 3.0 fi. A crude bottom type classification of unknown accuracy emerged as a by-product of the investigation.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 1995
Accession Number
ADA317774

Entities

People

  • Juanita C. Sandidge
  • Ronald J. Holyer

Organizations

  • United States Naval Research Laboratory

Tags

DTIC Thesaurus Topics

  • Accuracy
  • Airborne
  • Artificial Intelligence Computing
  • Artificial Intelligence Software
  • Australia
  • Bathymetry
  • Classification
  • Databases
  • Digital Information
  • Errors
  • Neural Networks
  • Radiance
  • Spectrometers
  • Training
  • Upwelling

Fields of Study

  • Environmental science

Readers

  • Atmospheric Remote Sensing.
  • Coastal and Marine Engineering/Sediment Transport/Hydraulic Engineering
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Neural Networks