支持向量机云检测方法的稳定性分析研究  

ON ANALYSING THE STABILITY OF SVM-BASED CLOUD DETECTION

作　　者：向大芳[1] 

机构地区：[1]广东松山职业技术学院计算机系,广东韶关512126

出　　处：《计算机应用与软件》2013年第1期226-228,241,共4页Computer Applications and Software

摘　　要：首先阐述支持向量机方法的原理;在此基础上,运用支持向量机方法对MODIS数据进行云检测试验,试验结果表明支持向量机方法适合进行云检测;最后分别从带宽变化和惩罚参数变化两个角度来分析支持向量机云检测方法的稳定性,结果表明支持向量机算法能很好地进行云检测,并具有较高的稳定性。Firstly, the principle of support vector machine （SVM） method is introduced in the paper. Based on this, the SVM method is applied to MODIS data for test of cloud detection, the testing result proves that the SVM method suits the cloud detection. At last, the stabili- ty of SVM-based cloud detection is analysed in terms of bandwidth alteration and penalty parameters alteration respectively. Result suggests that the SVM algorithm can be well used in cloud detection with high stability.

关 键 词：支持向量机 云检测 稳定性 

分 类 号：TP302.7[自动化与计算机技术—计算机系统结构]