New Methods of Change Detection Using Multispectral Data

Abstract

This report discusses the final phase of a project to develop multispectral change detection methods with emphasis on human activities. Two new algorithms were developed and tested: a Spectral/Spatial Classifier and a Feature Vector Spectral Classifier. The Spectral/Spatial method appears to be a powerful new tool. The Feature Vector results were inconclusive.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
May 15, 1991
Accession Number
ADA239370

Entities

People

  • Charles Sheffield
  • Gil Richardson

Tags

Communities of Interest

  • Air Platforms
  • C4I
  • Space

DTIC Thesaurus Topics

  • Abstracts
  • Algorithms
  • Artificial Intelligence
  • Change Detection
  • Computer Programs
  • Computers
  • Data Science
  • Data Sets
  • Detection
  • Equations
  • Information Science
  • Machine Learning
  • Materials
  • Multispectral
  • Statistics
  • Unsupervised Machine Learning
  • Word Processors

Readers

  • Computer Vision.
  • Systems Analysis and Design