一种基于融合核函数支持向量机的遥感图像分类

The remote sensing image classification based on fusion kernel function of support vector machine

从核函数选取规则着手,结合遥感数据本身特征,将具有互补性的几种核函数融合在一起,提出了一种复合核函数构造方法.通过实验数据与传统支持向量机方法比较,结果表明了复合核方法的有效性.

In the technology of remote sensing image classification,classification methods directly affect the classification accuracy between the samples.Current research of based on support vector machine（SVM）remote sensing image classification have achieved good results,but there is no further study on the selection of kernel function.From the selection of kernel function rules set out to research and combined with the feature of remote sensing data itself,fused several kinds of kernel function is complementary,put forward a method of composite kernel function,through compared with traditional SVM method,the experimental data show that the effectiveness of the composite kernel method.