论文部分内容阅读
为了提高云状自动识别的准确率,本文提出了一种新的云状自动识别方法.该方法基于湍流统计标度理论,将湍流的标度特征应用于云状的自动识别和分类中,对云图的灰度数据进行扩展自相似(ESS)模型标度分析,提取云图的标度指数特征,通过不同云系所得出的标度特征来识别云图.本文选用支持向量机作为分类器,对波状云、层状云、积状云、卷云和晴空五种云图进行识别.研究结果表明,通过提取ESS模型标度特征进行典型云状识别的准确率接近或超过90%.因此,基于湍流标度理论提取的云状特征是识别云状的显著性较强的特征,是对已有云状识别方法的一个很好的补充.
In order to improve the accuracy of cloud-like automatic recognition, a new cloud-like automatic identification method is proposed in this paper.The method is based on turbulent statistical scaling theory and applies the scaling features of turbulence to cloud-like automatic identification and classification. The gray scale data of the cloud graph is scaled by the extended self-similar (ESS) model, and the cloud index features are extracted, and the clouds are identified by scaling features of different cloud systems.In this paper, the support vector machine is chosen as the classifier, Clouds, stratiform clouds, cumulus clouds, cirrus clouds and clear sky.The results show that the accuracy of typical cloud-like identification by extracting ESS model scaling features is close to or over 90% .Therefore, The cloud feature extracted by the degree theory is a distinctive feature of recognizing clouds and is a good complement to the existing cloud-like identification methods.