Share Email Print

Journal of Applied Remote Sensing

Comment on the paper Design and analysis of wide-field-of-view polarization imaging spectrometer
Format Member Price Non-Member Price
PDF $20.00 $25.00
cover GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free. Check Access

Paper Abstract

The development of new multispectral sensors with unique band settings is critical for mapping the spatial distribution of increaser vegetation species in disturbed rangelands. The objective of this study was to evaluate the potential of WorldView-2 imagery for spectral classification of four increaser species, namely Hyparrhenia hirta, Eragrostis curvula, Sporobolus africanus, and Aristida diffusa, in the Okhombe communal rangelands of South Africa. The 8-bands were extracted from the WorldView-2 image, and 24 of the most widely used vegetation indices in estimating grassland biophysical parameters were calculated. The random forest algorithm and forward variable method were applied to identify the optimal variables (WorldView-2 spectral bands, vegetation indices, and a combination of bands and indices) for classifying the species. The random forest algorithm could classify species with an overall accuracy of 82.6% (<b<KHAT[an estimate of κ]=0.76</b< ) using six of the WorldView-2 spectral bands and an overall accuracy of 90% (<b<KHAT=0.87 </b<) using a subset of vegetation indices (=n<b<9</b< ). Three bands selected were located at the new WorldView-2 spectral regions of coastal blue, yellow, and the red-edge. There was no significant improvement in increaser species classification by using a combination of bands and indices. Overall, the study demonstrated the potential of the WorldView-2 data for improving increaser separability at species level.

Paper Details

Date Published: 18 September 2012
PDF: 17 pages
J. Appl. Remote Sens. 6(1) 063558 doi: 10.1117/1.JRS.6.063558
Published in: Journal of Applied Remote Sensing Volume 6, Issue 1
Show Author Affiliations
Khalid M. Mansour, Univ. of KwaZulu-Natal (South Africa)
Onisimo Mutanga, Univ. of KwaZulu-Natal (South Africa)

© SPIE. Terms of Use
Back to Top