Abstract: Cloud-type products have very important applications in many fields such as weather forecasting, climate change, disaster monitoring and early warning. In recent years, with the development of satellite technology, optical satellite remote sensing has become an important data source for cloud classification research due to its advantages of wide spatial coverage, high observation frequency and high spatial resolution. At the same time, cloud classification based on optical satellite remote sensing has also achieved a series of research results. In this paper, optical remote sensing satellites and their sensors for cloud classification are briefly introduced firstly. Then, the research progress of cloud classification based on optical satellite remote sensing in China and abroad in the past 20 years is systematically expounded. According to the principle and complexity of methods, cloud classification methods are mainly divided into three types based on simple linear relationship, mathematical statistics and machine learning, respectively. Finally, the problems existing in the research of cloud classification based on optical satellite remote sensing are analyzed and the future research direction is prospected, which is expected to provide reference for the research and application of domestic satellite cloud classification.

Key words: optical remote sensing satellite, cloud classification, mathematical statistics, machine learning