Logo image
Multi-temporal AVHRR digital data: an approach for landcover mapping of heterogeneous landscapes
Journal article   Peer reviewed

Multi-temporal AVHRR digital data: an approach for landcover mapping of heterogeneous landscapes

C G Fleischmann and Stephen J Walsh
Geocarto International, Vol.6(4), pp.5-20
1991
url
https://doi.org/10.1080/10106049109354336View
Published Version

Abstract

Geomatic Engineering
A multi-temporal sequence of seven NOAA-n, Advanced Very High Resolution Radiometer (AVHRR) satellite scenes (April 10, May 18, June 6, June 29, July 20, and August 18, 1987) were composited to derive cover-type information in the heterogeneous landscape of University Lake Watershed, North Carolina, U.S.A. The Normalized Difference Vegetation Index (NDVI) was calculated for each scene and merged into a seven-dimensional dataset, representing each time period sampled. An unsupervised classification was performed on the multi-temporal composite to derive five cover-type classes. Similar classifications were generated on single scene information. Ground control information was derived from an unsupervised classification of one kilometer grid compositional percentages initially derived from photo-interpreted landcover information. The multi-temporal NDVI classification more consistently characterized phenologic responses on a spatially dissected landscape than single scene clustering. Sub-pixel information showed how the algorithm separated compositional information between classes. Temporal vectors were plotted to illustrate differentiation on the basis of NDVI profiles.

Details

Metrics

612 Record Views
Logo image