Hyperspectral Imagery Classification Using a Backpropagation Neural Network

Abstract

A backpropagation neural network was developed and implemented for classifying AVIRIS (Airborne Visible/Infrared Imaging Spectrometer) hyperspectral imagery. It is a fully interconnected linkage of three layers of neural network. Fifty input layer neurons take in signals from Bands 41 to 90 of the AVIRIS spectral data in parallel. Test images are classified into four terrain categories of water, grassland, golf courses and built-up areas using four output neurons. A hidden layer consisting of 12 neurons is used. A training set containing 1,700 pixels for each of the four desired terrain categories is extracted and created from the first test image. Good classification accuracies of 81.8 percent to 95.5 percent are achieved despite the moderate AVIRIS pixel resolution of 20 meters by 20 meters. Backpropagation neural network, Hyperspectral imagery

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1993
Accession Number
ADA273615

Entities

People

  • Pi-fuay Chen
  • Tho C. Tran

Tags

DTIC Thesaurus Topics

  • Accuracy
  • Availability
  • Carbon Dioxide
  • Classification
  • Computer Programs
  • Computers
  • Computing Devices
  • Engineering
  • Hyperspectral Imagery
  • Interdisciplinary Science
  • Linear Accelerators
  • Military Applications
  • Neural Networks
  • Remote Sensing
  • Training
  • Two Dimensional
  • User Interface

Readers

  • Atmospheric Remote Sensing.
  • Neural Network Machine Learning.

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks