Image from Google Jackets

Evaluation of airborne multispectral imagery for mapping wetlands habitats and vegetation communities : Camp Lake assessment site, Alberta / Douglas R. Connery, Edward A. Cloutis, Tomasz Sankowski, Jim Devries, David Howerter, and Frances J. Dover.

By: Contributor(s): Material type: TextTextSeries: Proceedings of the 26th symposium on remote sensing of environment and the 18th symposium of the Canadian Remote Sensing Society. Information tools for sustainable development 375-378 Publication details: 1996Description: 28 cmLOC classification:
  • CON
Online resources: Summary: Airborne multispectral optical imagery was acquired over the Camp Lake Assessment Site in east central Alberta on August 4, 1995 to evaluate the applications of this imagery to mapping waterfowl habitats and vegetation communities. Two spatial resolution image sets were acquired: I.S x 1.5 m and 3 x 3 m. The imagery was classified using linear discriminant analysis and maximum likelihood classification in conjunction with ground ba~d data. The data suggests that both resolutions can provide comparabJe classiftcation accuracies. Classification accuracies are a function of the number of spectral bands, statistical analysis parameters (clas.'i thresholds), and image mtering. The results also suggest that the land cover classes used in the analysis exhibit significant spectral variability. Degradation of the 12 bit data to 8 bits has little effect on overall classification accuracies.

Includes bibliographical references (page 378).

Airborne multispectral optical imagery was acquired
over the Camp Lake Assessment Site in east central Alberta on
August 4, 1995 to evaluate the applications of this imagery to
mapping waterfowl habitats and vegetation communities. Two
spatial resolution image sets were acquired: I.S x 1.5 m and 3
x 3 m. The imagery was classified using linear discriminant
analysis and maximum likelihood classification in conjunction
with ground ba~d data. The data suggests that both
resolutions can provide comparabJe classiftcation accuracies.
Classification accuracies are a function of the number of
spectral bands, statistical analysis parameters (clas.'i thresholds),
and image mtering. The results also suggest that the land cover
classes used in the analysis exhibit significant spectral
variability. Degradation of the 12 bit data to 8 bits has little
effect on overall classification accuracies.

Ducks Unlimited Canada Institute for Wetland & Waterfowl (IWWR) Research Library, P.O. Box 1160, Stonewall, MB R0C 2Z0
(204)467-3276|Fax (204) 467-9028|