A radiometric preprocessing method of KOMPSAT-2 images for water quality monitoring

초록

Water quality monitoring for inland water from satellite images is an important issue as they offer periodical monitoring over wide area. Many researches have tried to measure water quality from low resolution satellite images such as LANDSAT images. High-resolution satellite images have a potential for monitoring spatial variation of water quality at finer scale. However, they have very small field-of-view and a sophisticated onboard processing. They typically have low signal to noise ratio over inland water and may show radiometric errors such as non-uniformity and non-linearity. These radiometric errors are caused by various elements in acquisition and transmission process of satellite images. These radiometric errors lead to increase uncertainty of water quality measurements. Therefore it is essential to check whether high resolution satellite images are suitable for water quality monitoring and to develop appropriate radiometric correction method. In this paper, we propose a radiometric preprocessing method of KOMSAT-2 images for water quality monitoring. We compare DN value distribution patterns of Kompsat-2 images with IKONOS and GEOEYE-2 images obtained at a similar time over inland water and discuss possible radiometric errors. We match DN values of a KOMPSAT-2 image against reference images obtained at a similar time. Using in-situ data and images corrected by the proposed method, we expect to extract water quality parameters such as concentrations of chlorophyll-a for water quality measurements.