Please use this identifier to cite or link to this item:
Title: Seasonal Variation in Land-Cover Classification Accuracy in a Diverse Region
Contributor(s): Sinha, Priyakant (author); Kumar, Lalit  (author)orcid ; Reid, Nick  (author)orcid 
Publication Date: 2012
DOI: 10.14358/PERS.78.3.271
Handle Link:
Abstract: This study determined the effect of seasonal spectral variability on land-cover classification of Landsat™ by comparing accuracies in different seasons in a mid-latitudinal (29°30' to 31°0'S) region with summer and winter rainfall, a broad altitudinal range, a temperate to subtropical climate and diverse land uses (e.g. summer and winter crops and nature conservation). The TC, PCA, and NDVI images were used alone and in combination with original bands for classification. Results using PCI-3, Bands 1 through 5, Bands 1 through 4, and 7 combinations were better than others, but the Bands 1 through 4 combination performed best across all dates. The highest accuracy was obtained with a mid summer (January) image (96.5 percent, Kappa 0.96) and the lowest with an early spring (September) image (86.7 percent, Kappa 0.84). The superiority of mid-summer images in generating highly accurate land-cover classifications was demonstrated by analysis of additional images under varying rainfall conditions in different years.
Publication Type: Journal Article
Source of Publication: Photogrammetric Engineering and Remote Sensing, 78(3), p. 271-280
Publisher: American Society for Photogrammetry and Remote Sensing
Place of Publication: United States of America
ISSN: 0099-1112
Field of Research (FOR): 090905 Photogrammetry and Remote Sensing
Socio-Economic Outcome Codes: 960504 Ecosystem Assessment and Management of Farmland, Arable Cropland and Permanent Cropland Environments
Peer Reviewed: Yes
HERDC Category Description: C1 Refereed Article in a Scholarly Journal
Statistics to Oct 2018: Visitors: 280
Views: 280
Downloads: 0
Appears in Collections:Journal Article
School of Environmental and Rural Science

Files in This Item:
2 files
File Description SizeFormat 
Show full item record


checked on Nov 30, 2018

Page view(s)

checked on Mar 1, 2019
Google Media

Google ScholarTM



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