Interactive Digital Image Processing Investigation. Phase II.

Abstract

The objective of the second phase of this investigation was to continue the development of the interactive multi-channel image classification capabilities of the DIAL system. This development proceeded in four directions. Formal demonstrations and a 'hands on' course in the DIAL algorithms implemented under the first phase of the investigation were given. Additional DIAL algorithms to support classification were developed, coded, and tested. These included a Program Module (PM) to apply the Karhunen - Loeve transformation to a multi-channel image, which has the effect of reducing the dimensionality of an image without significantly decreasing its information content. In addition two algorithms in refining class assignment by relaxation methods were developed. One was selected, then coded on DIAL and was applied to a classification of a LACIE intensive site, where it removed 'speckle', sharpened field boundaries, and increased the overall classification accuracy. A task to program the computationally intensive part of the maximum likelihood method on the STARAN was undertaken jointly with ETL. Finally an experiment in the maximum likelihood classification of a LANDSAT scene using the DIAL PMs was performed in cooperation with an ETL botanist. This experiment demonstrated the utility of the interactive classification algorithms in the study of the relationship between flora and geological structures. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Apr 01, 1980
Accession Number
ADA087518

Entities

People

  • J. S. Shipman
  • R. J. Spieler
  • W. C. Rice

Organizations

  • International Business Machines Corporation (Armonk, NY)

Tags

Communities of Interest

  • Sensors
  • Space

DTIC Thesaurus Topics

  • Computer Programming
  • Computers
  • Databases
  • Detectors
  • Digital Image Processing
  • Digital Images
  • Dimensionality Reduction
  • Feature Extraction
  • Image Processing
  • Information Science
  • Machine Learning
  • Materials
  • Numbers
  • Pattern Recognition
  • Plants
  • Supervised Machine Learning
  • Two Dimensional

Readers

  • Computer Vision.
  • Image Processing and Computer Vision.
  • Software Engineering

Technology Areas

  • AI & ML