Vegetation mapping of the Great Fish River basin, South Africa: Integrating spatial and multi‐spectral remote sensing techniques
- 24 February 2000
- journal article
- Published by Wiley in Applied Vegetation Science
- Vol. 3 (2) , 197-204
Abstract
We present a remote sensing based vegetation mapping technique well suited to a heterogeneous, semi‐arid environment. 10 structural vegetation classes were identified and described on the ground. Using Landsat‐TM from two different seasons and a combination of three conventional classification techniques (including a multi‐temporal classification) we were unsuccessful in delineating all of the desired vegetation classes. We then employed a simple tex‐tural classification index, known as the Moving Standard Deviation Index (MSDI), that has been used to map degradation status. MSDI measures spatial variations in the landscape and is calculated by passing a 3 × 3 standard deviation filter across the Landsat‐TM red band. High MSDI values are associated with degraded or disturbed rangelands whilst low MSDI values are associated with undisturbed rangeland. A combination of two conventional multi‐spectral techniques and MSDI were used to produce a final vegetation classification at an accuracy of 84 %. MSDI successfully discriminated between two contrasting vegetation types of identical spectral properties and significantly strengthened the accuracy of the classification. We recommend the use of a tex‐tural index such as MSDI to supplement conventional vegetation classification techniques in heterogeneous, semi‐arid or arid environments.Keywords
This publication has 21 references indexed in Scilit:
- A description of rangeland on commercial and communal land, Peddie district, South AfricaAfrican Journal of Range & Forage Science, 1999
- Textural Analysis of Historical Aerial Photography to Characterize Woody Plant Encroachment in South African SavannaRemote Sensing of Environment, 1998
- Changes to the vegetation of the mid‐Fish River valley, Eastern Cape, South Africa, in response to land‐use, as revealed by a direct gradient analysisAfrican Journal of Range & Forage Science, 1997
- Texture classification in aerial photographs and satellite dataInternational Journal of Remote Sensing, 1992
- On the accuracy of the global positioning system—A test using a hand-held receiverInternational Journal of Remote Sensing, 1992
- Soil and landscape processes in the restoration of rangelands.The Rangeland Journal, 1990
- Biological Feedbacks in Global DesertificationScience, 1990
- Remote Sensing Technology for Rangeland Management ApplicationsJournal of Range Management, 1989
- Land system mapping using landsat multispectral scanner imagery in the Willandra lakes region of western New South WalesAustralian Geographer, 1989
- Texture analysis of aerial photographsPattern Recognition, 1983