TY - JOUR
AU - Σκιάνης, Γ. Αιμ.
AU - Βαϊόπουλος, Δ.
AU - Νικολακόπουλος, Κ.
PY - 2004/01/01
Y2 - 2023/02/06
TI - ASSESMENT OF THE TVI VEGETATION INDEX WITH THE AID OF PROBABILITY THEORY
JF - Bulletin of the Geological Society of Greece
JA - geosociety
VL - 36
IS - 3
SE - GIS-Geoinformatics
DO - 10.12681/bgsg.16478
UR - https://ejournals.epublishing.ekt.gr/index.php/geosociety/article/view/16478
SP - 1338-1346
AB - <p>In the present paper the statistical behaviour of the Transformed Vegetation Index TVI is studied. TVI is defined by: (equation No1) - or, alternatively, by: (equation No2) <em>u </em>is the numerical value of the vegetation index, χ <em>and y </em>are the brightness values of the near infrared and red zones, respectively. Relation (1) defines the vegetation index TVI. Relation (2) defines the vegetation index TVI'. Using appropriate distributions to describe the histograms of χ and y channels, and taking into account certain theorems from probability theory, the expressions for the distributions of TVI and TVI' values are deduced. According to these expressions, the standard deviation of TVI image is larger than that of TVI', as well as NDVI (Normalized Difference Vegetation Index). The prevailing value of the TVI' histogram is located at the right part of the tonality range. Therefore, according to the mathematical analysis, the TVI image has a better contrast than that of the NDVI and TVI' images. The TVI' has a diffuse luminance. The theoretical predictions were tested with a Landsat 7 ETM image of Zakynthos Island (western Greece) and they were found to be in accordance with the satellite data. It was also observed that lineaments with a dark tonality are expressed more clearly in the TVI image than in the TVI' image. The general conclusion is that the TVI vegetation index is preferable from TVI', since the former produces images with a larger standard deviation and a better contrast than the latter. The results and conclusions of this paper may be useful in geological and environmental research , for mapping regions with a different vegetation cover.</p>
ER -