气象卫星云图统计特性的研究

A study on the statistical properties of the meteorological satellite cloud image

本文研究了气象卫星云图的灰度分布、信息熵、冗余度、二维傅里叶谱、一维及二维自相关函数以及卫星云图在差分域中的一些统计特征。结果表明,卫星云图的一维自相关函数不满足负指数单调下降的规律,不能用一阶平稳马尔科夫过程描述;云图的灰度分布具有一定的空间相关性,其空间频谱主要集中在空间频率较低的区域,差分图象的相关性明显减弱,但仍具有一定的冗余度,因此可以通过选择适当的方法对差分图象进行压缩编码而不丢失图象的信息。

Some statistical properties of the meteorological satellite cloud image (MSCI) such as image gray distribution, information entropy, redundancy, two-dimentional Fourier spectrum, one-and two-dimentional autocorrelation functions of the original MSCI and the gray distribution, the information entropy etc, in differential domain have been studied in this paper. It shows that the gray distribution of MSCI is not completely random but is somewhat spatially correlative, whose spatial spectrum energy is mainly distributed in the low frequency area and the spatial correlation is rather looser in the differential domain but the redundancy still exists, thus producing a compression encoding method for MSCI, which could keep the information from losing.