The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences
Download
Publications Copernicus
Download
Citation
Articles | Volume XLII-3/W2
https://doi.org/10.5194/isprs-archives-XLII-3-W2-131-2017
https://doi.org/10.5194/isprs-archives-XLII-3-W2-131-2017
16 Nov 2017
 | 16 Nov 2017

COMPARISON OF SPOT AND LANDSAT DATA IN CLASSIFYING WETLAND VEGETATION TYPES

M. T. Mosime and S. G. Tesfamichael

Keywords: Wetland vegetation, SPOT, Landsat, Unsupervised classification, Klipriviersberg Nature Reserve

Abstract. The aim of this study was to compare the performances of Landsat and SPOT imagery to map wetland vegetation types in the Klipsriviersberg Nature Reserve, South Africa. The Gauteng Conservation Plan 3.3 (C-Plan 3) was used to delineate the boundaries of the wetlands in the study area. According to the plan, the proposed study area falls within the Critical Biodiversity Areas (CBA) and Ecological Support Areas (ESA). Limited field data were collected within the boundaries of the wetlands during summer 2015 when the vegetation cover was relatively high. These data identified features including sparse vegetation, dense vegetation, grassland and bare land.Additional samples were added from Google Earth image to increase sample size. Both the field data and Google Earth data were used as reference against which the performances of SPOT and Landsat product were compared. Unsupervised classification was used to classify SPOT and Landsat images acquired in summer 2015. The results showed that overall accuracy of SPOT images is higher than Landsat images. This is attributed to its high spatial resolution of 1.5 m compared to 30 m spatial resolution of Landsat imagery. This indicates that SPOT imagery is recommended to map wetland vegetation diversity in a localised area such as the study area. The current high temporal resolution of the image has also an added advantage that conservationists should exploit.