基于支持向量机的EO-1 Hyperion遥感图像分类研究

以EO-1 Hyperion高光谱遥感数据为基础,对其进行大气校正、几何校正、滤波等处理,采用支持向量机分类方法对其进行分类,选择不同的核函数,主要有线性核、多项式核、径向基核、Sigmoid核,其余采用相同参数设置,进而比较不同核函数在EO-1 Hyperion数据分类中的效果。结果表明,采用支持向量机方法对研究区域的EO-1 Hyperion遥感数据进行分类,采用不同的核函数对分类结果影响不大。

EO-1 Hyperion Remote Sensing Image Classification Based on SVM

This research is based on the EO-1 Hyperion hyperspectral remote sensing data,the atmospheric correction,geometric correction, filtering and other processing were conducted,using support vector machine classification method to classify,choosing different kernel function,mainly including linear kernel,polynomial kernel,radial basis kernel,Sigmoid kernel,the rest of the same parameter setting,the effects of different kernel function in EO-1 Hyperion data classification were compared. The results show that： the use of support vector machine classification method to the research area of EO-1 Hyperion remote sensing data,using different kernel function has little effect on the classification results.