Assessment of vegetation indices using remote sensing (Case Study: Karaj- Iran)

Gathering information about the continuous changes of vegetation by conventional methods is difficult and expensive. In this case, usage of satellite will provide the possibility of extensive study of vegetation. The aim of this study is to evaluate the 5 vegetation indices in Karaj. For this study Landsat TM in the 1st of July 2013 was used. In order to achieve better results correcting images by using COS (t) in terms of atmospheric correction was done. Then NDVI, RVI, DVI, SAVI and TSAVI indices applied to the images then by using algorithm maximum likelihood classified into three classes poor coverage, medium coverage and no coverage. In order to assess the accuracy of maps, Error Matrix Analysis was used. overall accuracy and kappa coefficient was calculated for each index. Because overall accuracy can’t reflect the accuracy of map. So to reduce the impact of chance, Kappa coefficient was used. The results showed that NDVI index with the highest overall accuracy and Kappa coefficient 91/19, 87 best performances and SAVI with the least overall accuracy and Kappa coefficient 72/32, 68/4 has the weakest results among other indices. Which may be due to L. Because SAVI must be calculated correctly that should factor L moderating effect be calculated on optimum soil accurately and this requires knowledge of vegetation density that is not usually available.

Assessment of vegetation indices using remote sensing (Case Study: Karaj- Iran) Keywords:

Vegetation Index , satellite , kappa coefficient , overall accuracy. maximum likelihood