与传统最大似然（ML）分类器相比，支持向量机（SVM）在小训练样本时仍具有良好的分类性能，目前已广泛应用于多个领域。该文在极化SAR特征提取的基础上，将SVM应用于极化SAR图像分类，分析了分类器参数对分类性能的影响。利用NASA／JPL实验室AIRSAR系统的L波段旧金山全极化SAR数据比较了SVM和ML的分类性能，并进一步给出了基于SVM的国内某地区双极化SAR图像分类结果。

Compared with conventional maximum likelihood(ML) classifier,classification performance of support vector machine(SVM) is still good with small training samples.SVM is widely used in several fields recently.In this paper,SVM is used in classification of polarimetric SAR images based on feature extraction,and effect of several important parameters of SVM on classification performance is analyzed.The performance of SVM is compared with ML using L-band fully polarimetric SAR data of san francisco,acquired by the NASA/JPL AIRSAR sensor.A classification map of dual-polarization data,obtained by China,based on SVM is also presented.

TN957.52 TN958

感谢华东电子工程研究所张长耀主任提供国内双极化SAR数据.