Separation of Different Vegetation Types in ASTER and Landsat Satellite Images Using Satellite‐derived Vegetation Indices

Abstract

Separation of different vegetation types in satellite images is a critical issue in remote sensing. This is because of the close reflectance between different vegetation types that it makes difficult segregation of them in satellite images. In this study, to facilitate this problem, different satellite derived vegetation indices including: Normalized Difference Vegetation Index (NDVI), Enhanced Vegetation Index (EVI), and Enhanced Vegetation Index 2 (EVI2) were derived from the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) and Landsat-5 TM data. The obtained NDVI, EVI, and EVI2 images were then analyzed and interpreted in order to evaluate their effectiveness to discriminate rice and citrus fields from ASTER and Landsat data. In doing so, the Density Slicing (DS) classification technique followed by the trial and error method was implemented. The results indicated that the accuracies of ASTER NDVI and ASTER EVI2 for citrus mapping are about 75% and 65%, while the accuracies of Landsat NDVI and Landsat EVI for rice mapping are about 60% and 65%, respectively. The achieved results demonstrated higher performance of ASTER NDVI for citrus mapping and Landsat EVI for rice mapping. The study concluded that it is difficult to detect and map rice fields from satellite images using satellite-derived indices with high accuracy. However, the citrus fields can be mapped with the higher accuracy using satellite-derived indices.