针对Contourlet的多尺度、局部化、方向性和各向异性等优点,结合隐马尔科夫树(hidden Markov tree,HMT)模型和D-S(Dempster-Shafer)证据理论,提出一种新的SAR图像分割算法。该算法首先将隐马尔科夫树模型推广到Contourlet域,在多尺度HMT上采用D-S证据融合理论有效地融合Contourlet系数的持续性和聚集性,最后导出融合后的最大后验多尺度分割公式。本文算法对实测SAR图像进行分割试验,试验结果表明:与小波域上的HMT-MRF(Markovrandom field,MRF)融合分割及Contourlet域上HMT和MRF分割算法相比,本文算法在抑制斑点噪声的同时,可有效地提高SAR图像的分割精度。 更多还原
【Abstract】 Considering the Contourlet’s advantages of multiscale,localization,directionality and anisotropy,a new segmentation algorithm is proposed for SAR images based on hidden Markov tree in the Contourlet domain and D-S theory of evidence.Firstly,the algorithm extends the hidden Markov tree framework to the Contourlet domain.Then,the clustering and persistence of the Contourlet transform are effectively fused using D-S theory in HMT model,and the maximum a posterior(MAP) segmentation equation for the ... 更多还原