基于MARMA模型的SAR图像SVM分割

Support vector machine segmentation of SAR images based on MARMA model

在分析SAR图像特征的基础上,提出一种新的基于多尺度自回归滑动平均(multiscale autoregressive moving average,MARMA)模型的SAR图像分割方法.首先建立多尺度序列,然后通过研究SAR纹理图像的MARMA模型,建立适合SAR图像的多尺度特征矢量,最后采用提出的广义加权支持向量机进行特征分类.实验结果表明,采用此分割方法可以获得很好的分割结果.

According to the characteristics of SAR imagery, the support vector machine segmentation of SAR images was proposed based on multiscale autoregressive moving average （MARMA） model, which can capture the statistical scale-dependency of SAR images. Firstly, the multiscale sequences of SAR image were constructed. Secondly, methods for establishing MARMA model and extracting the multiscale stochastic characteristics of different SAR texture images were investigated. Finally, the characteristic vectors were classified using generalized weighted SVM. Experiments show the efficiency of the proposed algorithm.