论文部分内容阅读
在极化SAR影像极化特征的基础上,引入影像的纹理信息,利用带核函数的SSVM算法对极化SAR影像进行分类研究。该方法首先利用精致LEE滤波器对极化SAR影像进行去噪处理;然后采用小波变换对去噪后的总功率影像Span进行纹理特征提取;最后将纹理信息和极化信息结合,并采用SSVM方法对极化SAR影像进行分类。利用NASA/JPL AIRSAR获取的L波段SanFrancisco海湾和荷兰中部Flevoland地区的影像对该方法进行验证,结果表明,SSVM算法可有效地用于极化SAR影像分类,且分类精度和分类效率都优于SVM算法。同时纹理信息的引入使SSVM算法的分类精度得到了进一步提高。
Based on the polarimetric SAR image polarization features, texture information of the image is introduced, and the SSVM algorithm with kernel function is used to classify the polarized SAR images. Firstly, the refined LEE filter is used to denoise the polarimetric SAR image. Then, the wavelet transform is used to extract the texture features of the de-noised total power image Span. Finally, the texture information and polarization information are combined and the SSVM method is used Polarized SAR images are classified. The proposed method is validated using the L band SanFrancisco Bay and the Flevoland region of the Netherlands collected by NASA / JPL AIRSAR. The results show that the SSVM algorithm can be effectively used for polarimetric SAR image classification with better classification accuracy and classification efficiency than SVM algorithm. At the same time, the introduction of texture information makes SSVM algorithm classification accuracy has been further improved.