The effect of image radiometric correction on the accuracy of vegetation canopy density estimate using several Landsat-8 OLI's vegetation indices: A case study of Wonosari area, Indonesia

Abstract

Recent studies on the use of spectral indices have involved radiometric correction as a prerequisite. However, study on the effect of radiometric correction level on the accuracy of biophysical parameters' estimate is still rare in Indonesia. This study tried to investigate the influence of various radiometric correction levels and the number of vegetation strata on the accuracy of vegetation density estimates using NDVI, MSAVI2 and GEMI of Landsat 8 OLI. In this study, the dataset covering vegetated area in Wonosari, Gunung Kidul Regency, Indonesia was processed radiometrically using eight different methods, i.e. spectral radiance, at sensor reflectance, sun elevation correction, histogram adjustments using original DN, spectal radiance, at sensor reflectance, and sun position correction respectively, as well as dark object subtraction (DOS). Every image with specific correction level was then transformed using the aforementioned indices, in order correlate with the field-measured canopy density. The analysis were carried out by considering the number of canopy layers. This found that different radiometric correction methods resulted canopy density estimates with different accuracies. The number of canopy strata also played an important role. Every vegetation index transformation performed its best accuracy by using different radiometric correction method and different number of canopy layers.