SAR图像的多尺度建模与分割

Multiscale Modeling and Segmentation of SAR Image

根据合成孔径雷达图像的成像机理,基于遗传算法,提出了SAR图像的期望最大建模方法和多尺度无监督分类方法。利用最小长度准则能够有效地确定SAR图像分类数,且集遗传算法和EM算法的优点于一身,使得算法能够取得全局最优结果。试验结果表明:该分割方法是可行的,与其它方法相比,分割质量有明显改进。

A modeling method with maximum expectation and multiscale unsupervised classification method are proposed based on the genetic algorithm （GA） and imagery mechanism of synthetic aperture radar （SAR） imagery. The algorithm is capable of selecting the number of classification of SAR images using the minimum description length （MDL） criterion for Gaussian mixture model. Our approach benefits from the properties of genetic algorithms （GA） and the EM algorithm by combination of both into a single procedure. Some experimental results are given based on our proposed approach and show that the GA-EM outperforms the other methods.