基于加权合成核与三重Markov场的极化SAR图像分类方法

马尔可夫随机场(Markov Random Field,MRF)广泛用于处理遥感图像的分类问题,然而MRF在构建极化合成孔径雷达(Synthetic Aperture Radar,SAR)图像模型时未考虑其非平稳特性且对初始分类较为敏感,为此本文提出了一种基于加权合成核与三重马尔可夫随机场(Triplet Markov Field,TMF)的极化SAR图像分类方法.该方法依据训练样本在特征空间上的距离,提出了加权合成核函数权重系数的自适应确定方法以提高初始分类的精度和普适性;为充分考虑极化SAR图像的非平稳统计特性,利用TMF对极化SAR图像进行统计建模以实现贝叶斯分类.实验结果表明,与基于MRF的极化SAR图像分类方法相比,本文所提方法可获得更高的分类精度和更平滑的同质区域分类结果,而且本文方法能更好地保持图像边缘信息.

A Classification Method of PolSAR Image Based on Weighted Composite Kernel and Triplet Markov Field

Markov random field (MRF) is widely applied to remote sensing images classification.However,the MRF-based classification method does not take the nonstationarity properties of images into account when it models polarimetric syn-thetic aperture radar ( PolSAR) images,and is sensitive to the initial classification.Therefore,this paper proposes a classifica-tion method of PolSAR image based on the weighted composite kernel and the triplet Markov field ( TMF) .Based on the dis-tances between the features of training samples,we compute the kernel weights of the weighted composite kernel for improving the accuracy and popularity of the initial classification.Then,taking the nonstationarity properties of PolSAR images into con-sideration,the TMF is introduced to model the statistics of real PolSAR images to realize the Bayesian classification.Experi-ments indicate that the proposed method can obtain higher classification accuracy and smoother homogeneous areas than the MRF-based PolSAR image classification method.Moreover,the proposed method can get more accurate edge location.