Color compensation for cloud shadows

Cloud covers are generally present in optical satellite images, thereby limiting the usage of optical images and increasing the difficulty of image analysis. In this paper, a cloud shadow compensation approach aiming at easing difficulties suffered from cloud covers is introduced. In the proposed approach, information of shadow pixels is compensated by a statistic adjustment method and a global optimization process. Experimental analyses on the satellite images acquired by the Landsat-7 Enhanced Thematic Mapper Plus (ETM+) sensor were conducted. The results including quantitative and qualitative analyses on simulated data show that our method can successfully compensate for the intensities of shadow pixels.