Proceedings Article | 12 October 2011
KEYWORDS: Vegetation, Earth observing sensors, Solar radiation, High resolution satellite images, Satellites, Satellite imaging, Reflectivity, Sensors, Light sources and illumination, Sun
A satellite image, which periodically records the spectroscopic characteristics of objects, is used to calculate the index of
vegetation.. However, it can have different results according to sensor correction, the state of atmosphere, geometric
relations between the sun, sensor and object, and a diversity of other factors. So although it is a picture of the same area,
if it was taken at a different time, it is difficult to expect the same value of lightness. This paper converses DN values of
KOMPSAT-2 and IKONOS, satellites with high resolution, into Reflectence value and then calculates NDVI and TCTGVI.
It uses the calculated index of vegetation and classifies between vegetation areas and non-vegetation areas to
conduct a comparative analysis of the rate of concordance. And it calculates the solar radiant quantities to conduct a
comparison of correlations with the areas where vegetation is wrongly classified. The analysis reveals that the rate of
concordance in the areas of vegetation is high: NDVI 85%, GVI 85.35%. It shows a similar rate of concordance in NDVI
and GVI. And the solar radiant quantities on average in all the researched areas are large: KOMPSAT-2 207.5w/m2 and
IKONOS 668.4w/m2 those in the wrongly classified areas are small: KOMPSAT-2 128.2w/m2 and IKONOS 553.7w/m2.
The reason that the wrongly classified areas have small solar radiant quantities is that they have a shadowed part
influenced by its topography, it is judged. In particular, in the winter, when the solar radiant quantities are comparatively
small, KOMPSAT-2 has difficulty collecting vegetation, not more than 128.175 w/m2, which means, it is judged, that it
is impossible to normally detect the Red band and the NIR band which are used for calculating the index of vegetation
from a satellite.
