Application of Geostatistical Methods and Wavelets to the Analysis of Hyperspectral Imagery and the Testing of a Moving Variogram

Abstract

Part I of this report describes the analysis on moving averages, variances and variograms for NIR from a SPOT image of part of Fort A. P. Hill. The relation between elevation from a digital elevation model, the raw elevation data and NIR and NDVI were explored in detail. The correlations are weak in spite of an apparent visual relation. Part II describes the analyses of the hyperspectral hymap' data. The results illustrate the difficulty of deciding how many and which wavebands to retain. Certain groups of bands reappear in different analyses, but equally there are less stable groupings. The PCA results do not discriminate as well as the raw variograms and certain classifications. Eight groups of bands seem to appear regularly. The pixel maps show that even within these groups different information emerges. Part IV of the report describes the analysis of the National Soil Inventory of England and Wales. Geostatistical methods and wavelet analysis are compared. Part IV of the report of the project is an aide memoire for sampling. It embraces both design-based sampling, which is based on classical statistics, and model-based sampling which is underpinned by geostatistics. This work is a guide to sampling field-based information or pixels from images.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 23, 2001
Accession Number
ADA389183

Entities

People

  • Margaret A. Oliver

Organizations

  • University of Reading

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Agriculture
  • Computer Programs
  • Data Science
  • Data Sets
  • Databases
  • Digital Elevation Models
  • Ecology
  • Factor Analysis
  • Geographic Regions
  • Hyperspectral Imagery
  • Information Science
  • Measurement
  • Normal Distribution
  • Remote Sensing
  • Statistical Sampling
  • Statistics
  • Surveys

Readers

  • Business Analytics
  • Image Processing and Computer Vision.
  • Statistical inference.