Delineation of riparian vegetation from Landsat multi-temporal imagery using PCA
Autor: | Neil Sims, Olga Barron, Irina Emelyanova, Masoomeh Alaibakhsh, Alireza Mohyeddin, Mehdi Khiadani |
---|---|
Rok vydání: | 2016 |
Předmět: |
geography
geography.geographical_feature_category 010504 meteorology & atmospheric sciences Digital data 0211 other engineering and technologies 02 engineering and technology 01 natural sciences Thresholding Normalized Difference Vegetation Index Principal component analysis medicine Environmental science Satellite imagery medicine.symptom Scale (map) Vegetation (pathology) 021101 geological & geomatics engineering 0105 earth and related environmental sciences Water Science and Technology Remote sensing Riparian zone |
Zdroj: | Hydrological Processes. 31:800-810 |
ISSN: | 0885-6087 |
DOI: | 10.1002/hyp.11054 |
Popis: | 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 (PCA) applied to multi-temporal Normalised Difference Vegetation Index (NDVI) 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 NDVI 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. This article is protected by copyright. All rights reserved. |
Databáze: | OpenAIRE |
Externí odkaz: |