Investigating land cover effect on dark object subtraction model of atmospheric correction

An efficient and low-cost atmospheric correction technique is becoming more and more necessary as the fields requiring remote sensing imagery grow. Certain techniques rely on measurements taken on-site, while others rely merely on the image; the latter requires less resources and time. For atmospheric correction, one of the most widely used techniques is the Dark Object Subtraction (DOS). In this study, Landsat-8 and Sentinel-2 data are subjected to the DOS and cosine of the solar zenith correction COST algorithm under varying atmospheric circumstances (July and December). Four distinct land cover classes (water, vegetation, desert, and urban) are used to categorize the data. The corrected DOS scenes are compared using both qualitative and quantitative analysis to the Level 2A of each satellite, as well as to the outcomes of QUick Atmospheric Correction (QUAC) in the case of Landsat-8 and Sentinel-2 Correction (Sen2Cor) results. As the Sentinel-2 DOS demonstrates good results in the vegetation and water scene, the results demonstrate the effectiveness of the DOS in the case of Landsat-8 for the vegetation, water, and urban scenes. But because there were no dark objects in the desert environment, the results were unsatisfactory, especially in clear weather for both satellites