云的定量识别在卫星数据反演中非常重要。云检测结果的质量直接影响各种反演产品的准确性。云检测实际上是一种目标识别和分类的过程, 检测的目的是为了特征提取, 因此,大量信号与系统中的新兴算法都被运用于云检测的技术研究中来。匹配追踪算法是近年来发展起来的非常有效的特征提取算法, 而正交匹配追踪算法更能有效提高信噪比。因此文中利用正交匹配追踪算法与多通道阈值法相结合, 进行了卫星云图云检测的相关研究。通过MODIS的云检测试验表明, 将正交匹配追踪算法应用到多光谱云图的信息处理中可以有效提高云检测的精度。

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.