DSpace
 

KNUSTSpace >
Research Articles >
College of Engineering >

Please use this identifier to cite or link to this item: http://hdl.handle.net/123456789/10585

Title: Fast Fourier Transform (FFT) Analysis on NDVI Time Series for Assessment of Vegetation Patterns in Ghana
Authors: Quarshie, Gilbert Kobla
Osei Jnr., Edward Matthew
Prah, Benjamin E. Kwasi
Amoah, Adwoa Sarpong
Issue Date: Jan-2015
Publisher: International Journal of Remote Sensing & Geoscience (IJRSG)
Citation: International Journal of Remote Sensing & Geoscience (IJRSG); Volume 4, Issue 1, Jan. 2015
Abstract: Time series analysis of Normalized Difference Vegetation Index (NDVI) imagery is a powerful tool in studying vegetation phenology in data scarce and inaccessible areas. Application of these datasets involves typically, per-pixel analysis of multi-temporal vegetation indices (VIs), which are frequently subject to high-frequency fluctuations (i.e. noise) caused by changing atmospheric conditions and varying sun-sensor-surface geometries. A broad range of NDVI noise-reduction strategies are applied in an effort to either reduce or if possible remove completely such embedded noise in these dataset. In light of this, the Fast Fourier Transform (FFT), a noise reducing algorithm in Erdas Imagine was applied to two averaged NOAA - AVHRR NDVI images acquired from Global Inventory Modeling and Mapping Studies (GIMMS) for Africa. The two averaged images were calculated from 792 NDVI images representing 1996 (i.e. 1986 -1996, 396 images) and 2006 (i.e. 1996 -2006, 396 images). Supervised clustering and maximum likelihood classifications algorithm were used to classify the NDVI dataset and vegetation maps representing 1996 and 2006 produced. These maps were visually assessed for land cover changes during the 10 year period for Ghana, and the comparison shows an alarming 11.10% increase in Grassland / Bare Surface and Settlement most probably due to increase in population. Closed or Evergreen forest decreased by 0.39%, whereas Mixture of Closed and Open Forest increased by 5.95% due to conscious effort of reafforestation programs by the government.
Description: An article published by International Journal of Remote Sensing & Geoscience (IJRSG); Volume 4, Issue 1, Jan. 2015
URI: http://hdl.handle.net/123456789/10585
Appears in Collections:College of Engineering

Files in This Item:

File Description SizeFormat
FAST FOURIER TRANSFORM (FFT) ANALYSIS ON NDVI TIME SERIES FOR ASSESSMENT OF VEGETATION PATTERNS IN GHANA(1).pdf790.63 kBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Valid XHTML 1.0! DSpace Software Copyright © 2002-2010  Duraspace - Feedback