基于Contourlet域图谱聚类和多尺度Markov模型的多光谱遥感图像分割

针对多光谱遥感图像的特点,结合图谱聚类、Co ntourlet系数分布的统计特性和多尺度Markov模型, 提出了一种基于Contourlet域图谱聚类和多尺度Markov模型的分割(CSCMMS)方法。首先对 待分割图像进行Contourlet变换,利用图谱聚类对最粗尺度低频图像聚类得到可靠的初始分 割结果；然后 利用互信息构造Contourlet域的多尺度Markov模型,结合多尺度、多方向的图像信息将低频 图像的初始分 割结果逐尺度传递到最细尺度,得到原始图像的分割。对合成图像和多光谱遥感图像的实验 结果表明,提 出方法在边缘信息保持和噪声敏感性上具有明显改进,错分率和运算时间进一步降低。

Segmentation of multispectral remote sensing image based on spectral clustering and multiscale Markov model in Contourlet domain

Spectral clustering methods have recen tly shown great promise for the problem of image segmentation.However,the comput ational demands of these approaches make them infeasible to large problems such as multispectral remote sensing images.According to the particularities of multi spectral remote sensing images,a segmentation method is proposed based on spectr al clustering,statistics characteristics of Contourlet coefficients and multisca le Markov model.First,the spectral clustering is implemented to the coarsest low pass image of multispectral remote sensing image in Contourlet domain to obtain the initial segmentation result;Second,a multiscale Markov model,which contains the initial segmentation result as the coarsest scale,is constructed using mutua l information to capture the relationship between Contourlet coefficients in the same scale and across scales;Third,the final segmentation result is obtained by confusing multiscale and multi-directional image information based on the mult iscale Markov model.Compared with the classical spectral clustering method and t he multiscale segmentation method based on HMT model in Contourlet domain,the se gmentation results for both synthetic images and real multispectral remote sensi ng images show that the proposed method not only has better performance in edges preservation and noise sensitivity,but also has lower misclassification probabi lity and running time,and it can achieve satisfactory segmentation results.