多尺度马尔科夫随机场在基于高分辨率合成孔径雷达影像的城市地物分类中的应用

An Application of Urban SAR Images Classification Utilizing Multi-scale Markov Random Field

本文通过基于多尺度马尔科夫随机场模型的最大似然算法及基于传统马尔科夫随机场势函数的像素间Gabor相似方法,使用迭代条件模型对影像进行分割,使用K均值分类算法对分割后的影像进行分类,选择北京市通州区作为研究区,使用上述方法对多幅该区域的高分辨率合成孔径雷达影像进行了分类,新方法可以实现优于传统算法的分割精度,能够清晰区分建筑物之间的边界。

The authors obtained the initial segmentation using the Maximum likelihood（ML） algorithm based on the multi-scale Markov Random Field（MRF） model and involved the Gabor similarity between pixels based on the traditional MRF potential function, and employed the Iterative Conditional Model algorithm to implement the segmentation. And we classified the segmentation image by using the K-means classification algorithm. The experimental results on several real SAR images showed that the proposed approach performs better than traditional methods in the segmentation accuracy, and building boundaries were clearly obtained by the proposed approach.