Transformation of surface reflectance of Göktürk-2 images using image-based methods

Atmospheric correction is an important requirement in the pre-processing of satellite images for many applications. In this study, the validity of image-based atmospheric correction methods for Göktürk-2, the first high-resolution reconnaissance and surveillance satellite of Turkey, was evaluated. A comparison was made between the Top of Atmosphere (ToA), Dark Object Subtraction (DOS), and Cosine of the Solar Zenith Angle Correction (COST) methods applied to the Göktürk-2 image. First the image radiance value was calculated, and then the reflectance values were obtained using ToA, DOS and COST algorithms. Normalized Difference Vegetation Index (NDVI) was calculated to evaluate the quality of the methods. A cross-section corresponding to the same area was taken from the NDVI images and comparison was made between image-based atmospheric correction methods. In comparison, it was seen that DOS and COST methods eliminate atmospheric effects by using the information in Göktürk-2 images without any real-time measurements. COST method gives more accurate results in Göktürk-2 images since it is based on images and uses atmospheric transmittance information in the direction of illumination.