Premium
Delineation of riparian vegetation from Landsat multi‐temporal imagery using PCA
Author(s) -
Alaibakhsh Masoomeh,
Emelyanova Irina,
Barron Olga,
Sims Neil,
Khiadani Mehdi,
Mohyeddin Alireza
Publication year - 2016
Publication title -
hydrological processes
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.222
H-Index - 161
eISSN - 1099-1085
pISSN - 0885-6087
DOI - 10.1002/hyp.11054
Subject(s) - riparian zone , thresholding , vegetation (pathology) , remote sensing , environmental science , ground truth , normalized difference vegetation index , scale (ratio) , enhanced vegetation index , vegetation index , geology , geography , cartography , computer science , ecology , artificial intelligence , climate change , medicine , oceanography , pathology , habitat , image (mathematics) , biology
A deficiency in crucial digital data, such as vegetation cover, in remote regions is a challenging issue for water management and planning, especially for areas undergoing rapid development, such as mining in the Pilbara, Western Australia. This is particularly relevant to riparian vegetation, which provides important ecological services and, as such, requires regional protection. The objective of this research was to develop an approach to riparian vegetation mapping at a regional scale using remotely sensed data. The proposed method was based on principal component analysis applied to multi‐temporal Normalized Difference Vegetation Index datasets derived from Landsat TM 5 imagery. To delimit the spatial extent of riparian vegetation, a thresholding method was required and various thresholding algorithms were tested. The accuracy of results was estimated for various Normalized Difference Vegetation Index multi‐temporal datasets using available ground‐truth data. The combination of a 14‐dry‐date dataset and Kittler's thresholding method provided the most accurate delineation of riparian vegetation.