利用极化目标分解和WMRF的全极化SAR图像分类方法
Classification of Fully Polarimetric SAR Image Based on Polarimetric Target Decomposition and Wishart Markov Random Field
-
摘要: 提出了一种新的全极化SAR图像非监督分类方法,该方法将H/Alpha/A分解与马尔科夫随机场(Morkov rondom field,MRF)相结合。首先,根据地物的散射机制进行H/Alpha/A分解得到初始分类;然后,由基于Wishart分布的最大似然法迭代聚类更新分类结果;最后,结合WMRF(Wishart Markov randomfield)方法,由迭代条件模型法求取最大后验准则下的分割结果。NASA/JPL实验室的数据结果表明,该算法具有较好的分类效果,并获得了较高的分类精度。Abstract: A new classification method is proposed for polarimetric SAR images.The H/Alpha/A decomposition is combined with Markov random field model.Firstly,H/Alpha/A decomposition is adopted to obtain 16 initial clusters based on their scattering characteristics.Secondly,the wishart distribution based on maximum likelihood is implied to update the classification result.Finally,by the adoption of wishart Markov random field model,iterated conditional model method(ICM) based on maximum a posteriori criterion is used to acquire the final classification result.Using fully polarimetric SAR images acquired by the NASA/ JPL,the experimental results verify the effectiveness of this improved algorithm.
下载:
