All-sky Imaging Radiometer On The Foundation Of Cloud And Weather, Remote Sensing,

In this paper a system has been developed for studying characteristics of "clouds and sky" on the images taken by a whole-sky camera at three different wave bands (the visible, the infrared, long wave infrared), then assessing the presence, distribution, shape, and radiance of clouds over the entire sky using automated cloud decision algorithms and related processing. This study lays focuses on the visible-light spectrum, and it is composed as follow:In first part, the fundamental principles and theories of the digital image processing are presented.The second part shows the methods in detail of analyzing and processing all the images. We find that the sky pixel's blue /brightness (red, green and blue) ratio is always more than 0.375 and the cloud pixel's is less than 0.345 when the day is fine. But it is difficult to get rules of those images, which are taken in the condition of the occurrence with a great deal of aerosol in the sky. Such images are analyzed by employing the other methods in the study.Finally, the paper presents the results after processing images data. It includes estimating the cloud-cover, and describing cloud shape, analyzing radiation distribution of clear-sky. The estimation of cloud-cover is mainly affected by the visibility of the day. Comparing the calculated results from about 350 images to the observation, the average error is about 10%. We also compare the analysis of radiance to the one analyzed with a model, they fits very well. But we can't describe the radiance quantificationally at present. At the end of the paper, some differences between those images taken at three different places are analyzed.