Although the Cloud-Aerosol Lidar with Orthogonal Polarization (CALIPSO) has been widely used in aerosol research, the classification of aerosol and cloud still exist some problems. Tradition classification method used by NASA is probability distribution functions (PDFs), but in reality, when we want to realize this algorithm, we fund it is difficult to describe the multi-modal distribution of cloud backscatter coefficients. Further, because ice cloud and dust aerosol have some similar properties, so it is not easy to identify them. In this paper, we introduce a classification method which based on Support vector machine (SVM), and add another characteristic. Then according to the result of classification inverse the aerosol characteristic, the height of cloud top, at the same time, combine with the CloudSat calculate the other cloud character, these data will be helpful for further climate research.
展开▼
