Study on the cloud masking method of geostationary ocean color imager utilizing a model-based surface reflectance over land surface

GOCI is the world's first geostationary ocean color satellite. Although the main mission of GOCI is monitoring ocean phenomena surrounding the Korean peninsula, the sensor can be utilized for the land applications. Because GOCI is able to observe not only oceanic regions, but also East Asian land surface. To obtain land information accurately, cloud masking should be applied. In this study, we tried to develop the cloud masking algorithm concentrated on thin clouds over the land cover. Especially, BRDF modeling has performed to consider influences on the land surface. There are two steps in the cloud masking algorithm we suggested. The first step is composed of the simple reflectance threshold method using four channels of GOCI to decide thick cloud pixels. Next, the difference method is performed using computed differences between the surface reflectance based on BRDF modeling and TOA reflectance. As a result, probability of cloud detection is getting higher when using both steps of cloud masking algorithm.