【摘要】 聚类作为一种重要的图像分割方法得到了大量研究,提出了一种新的结合稀疏编码的红外图像聚类分割算法,扩展了传统的基于K-means聚类的图像分割方法。结合稀疏编码的聚类算法能有效融合图像的局部信息,而且易于利用像素之间的内在相关性,但是对于分割会出现过分割和像素难以归类的问题。为此,在字典的学习过程中,将原子的聚类算法引入其中,有助于缩减字典中原子所属类别的数目防止出现过分割;同时将稀疏编码系数同原子对聚类中心的隶属程度相结合来判断像素所属的类别。这种处理方式能更好地实现利用像素的内在相关性进行聚类分割,并在其中自然引入了局部空间信息,达到更好分离目标区域和背景区域的目的。实验结果表明,结合稀疏编码的K-means聚类分割算法能更好的实现复杂背景下红外图像重要区域的准确分割提取。更多还原

【Abstract】 Clustering is an important method for image segmentation,and has got much research.A new algorithm for infrared image segmentation based on clustering combined with sparse coding is proposed.The traditional image segmentation method based on K-means clustering is extended.The clustering algorithm combined with sparse coding can fuse the local information of image.The inner relationships of pixels are used.But it produces the problem of over-segmentation and difficulty in pixels classification for segmentation.The clustering method is introduced for atoms in dictionary learning.The class number of atoms in dictionary is reduced in order to avoid over-segmentation.The class of pixels is estimated by combining the sparse coefficients and the degrees of membership with the atoms to cluster center.The usage of inner relationships of pixels and the local information can help to enhance the segmentation performance of background and target area.The experimental result shows that the important area can be separated well under complex background in infrared image by this method.更多还原