Cloud detection algorithm based on the Orthogonal Matching Pursuit

Quantitative identification of cloud is very important in meteorological satellite data retrieval, and the result of cloud detection affects the accuracy directly. In fact, the cloud detecting technology is actually a process of distinguishing the objects and background, and the purpose of detection is to extract cloud features. Therefore, much signals processing and system algorithms have been applied to the technology of cloud detection. The matching pursuit algorithm (MP) is a very effective algorithm for feature extraction, which is developed in recent years, and the Orthogonal Matching Pursuit algorithm (OMP) can improve the signal-to-noise ratio more effectively. In this paper, Orthogonal Matching Pursuit algorithm and multi-channel threshold method were combined to carry out relevant research on cloud detection of MODIS data. Based on the MODIS cloud detection results, it could be proved that the integrative algorithm of multi-channel threshold combined with the Orthogonal Matching Pursuit algorithm would be more effective to cloud detection.