论文部分内容阅读
本文提出一种对极化合成孔径雷达(SAR)图像进行自动多分辨率分类的方法。首先利用多视极化白化滤波(MPWF)抑制极化SAR图像的相干斑,得到反映地物辐射特征的纹理SAR图像,然后利用小波变换(WT)提取不同分辨率的纹理信息,在最低分辨率级利用Akaik信息准则(AIC)自动估计图像中的纹理类数,进而在各个分辨率级利用马尔可夫随机场(MRF)模型表征各像素间的空间关联信息,并分别利用最大似然(ML)方法和循环条件模式(ICM)进行自动的模型参数估计和最大后验概率(MAP)分类,最后应用NASA/JPL机载L波段极化SAR数据验证了本文所提分类方法的有效性和优越性。
This paper presents a method of automatic multi-resolution classification of polarized synthetic aperture radar (SAR) images. Firstly, MPWF is used to suppress the speckle of SAR images, and the texture SAR images which reflect the radiation characteristics are obtained. Then the texture information of different resolutions is extracted by wavelet transform (WT) Level (Akaik’s information criterion (AIC)) is used to automatically estimate the number of texture classes in the image. Markov random field (MRF) model is then used to characterize the spatial correlation information of each pixel at each resolution level. Maximum likelihood ) Method and cyclic condition model (ICM) were used to automatically estimate the model parameters and classify the maximum posteriori probability (MAP). Finally, the validity and superiority of the classification method proposed in this paper were verified by using NASA / JPL airborne L-band polarimetric SAR data Sex.