暴雨过程的卫星云图纹理特征研究

Textural Features of Satellite Images in Rainstorm Process

针对卫星云图上云团所具有的半流体特性,应用纹理特征分析法对其进行定量研究。首先用灰度共生矩阵统计法提取100个样本云团的纹理特征参数,然后应用Mahalanobis距离作相似性测度进行聚类分析,提取出4类云以及无云区的具有旋转不变性的纹理特征参数。在此基础上分析了2003年7月4—5日江淮流域发生特大暴雨地区云团的纹理特征参数的变化规律与降水过程之间的对应关系。从实际的降水资料和纹理特征参数变化的回归分析结果来看,纹理特征参数的变化规律与降水量之间有较好的对应关系。这些结果不仅对中短期天气过程的诊断分析而且对数值预报都具有一定的实用意义。

The Grey Level Co-occurrence Matrix(GLCM) approach was firstly used in this paper to extract the textural parameters,the intrinsic attribute of different types of clouds,from 100 satellite images of cloud cluster samples,and then,four texture parameters of cirrus,cumulus,cumulonimbus,stratus and cloudless areas were clustered by using the Mahalanobis distance as the comparability measure standard.In the experiment,textural features of the clouds,of the strong rainstorm in the period of July 4-5,2003 in the middle and lower reaches of the Changjiang River,and their relations with precipitation were analyzed.Results of the quadratic polynomial regression analysis show that the developing tendency of texture parameters is parallel with one of precipitation.These results are of practical significance for the diagnoses and numerical prediction of weather processes,especially of short-medium range rainstorm processes.