基于多尺度随机模型的SAR图像无监督分割

合成孔径雷达（synthetic aperture radar，SAR）是一种基于相干原理的成像系统，在SAR图像中存在严重影响图像质量的斑点噪声，使得SAR图像的可靠分割非常困难。〖BP）〗根据SAR图像的成像机理，利用两种多尺度随机模型，即多尺度自回归（Multiscale Autoregressive，MAR）模型和多尺度自回归滑动平均（Multiscale Aautoregressive Moving Average， MARMA）模型，分别来描述同一场景不同分辨率SAR图像像素间的统计相关性，并构造了相应的多分辨混合算法实现SAR图像的无监督分割。试验结果表明，提出的两种无监督分割方法是可行的，且MARMA模型比MAR模型能够更精确地捕捉SAR图像多尺度序列中不同类型地形的统计信息，使分割质量具有明显的改进。

Unsupervised segmentation of SAR image based on multiscale stochastic model

The presence of speckle in Synthetic Aperture Radar(SAR) images makes the segmentation of such images difficult,either by gray levels or by texture.According to the mechanism of SAR imaging,two unsupervised segmentation methods were proposed based on two class of multiscale stochastic model,namely multiscale autoregressive(MAR) model and multiscale autoregressive moving average(MARMA) model.These models capture the statistical information in a multiscale sequence of SAR image,which is then used to implement unsupervised segmentation of SAR image via multiresolution mixture algorithm.Experimental results over SAR images confirm the proposed segmentation methods are valid.