中文摘要：

由于传统图像分割方法对噪声的敏感性和检测结果的不连续性等问题导致图像分割精度较低,提出一种基于多层马尔科夫随机场模型融合的图像分割方法。首先分别通过模糊C均值聚类（FCM）方法和马尔科夫随机场（MRF）方法得到两个分割效果较差的图像,随后运用多层马尔科夫随机场模型的融合特性将两个传统方法得到的分割结果进行融合。该方法运用多层马尔科夫随机场融合方法引入邻域内像素间相关性和各层间的联系,并且在实验中得出与两个传统方法相比较更细致和精确的结果。实验结果表明,多层马尔科夫随机场模型的融合方法可以将两个传统分割方法的结果较好地融合,并且得到更加精确的结果。

英文摘要：

Because of the sensitivity to noise and the discontinuity of detection results the traditional image segmentation method has,which result in low precision of image segmentation,this paper proposes an image segmentation method,it is based on multilayer Markov random field models fusion. First,we obtain two images with poorer segmentation results based on fuzzy c-means algorithm（ FCM） and Markov random field（ MRF） method separately. Then we use the fusion characteristic of multilayer MRF model to fuse these segmentation results derived from two traditional methods. The approach proposed in this paper makes use of the multilayer MRF fusion method to introduce the correlation between the pixels within neighbourhood and the links of each layer,and derives from experiment the more detailed and precise results compared with two traditional methods. Experimental results indicates that the fusion method of multilayer Markov random field model can better fuse the results of traditional segmentation methods,and achieves a more precise result.