基于多层二部图的高光谱模糊聚类算法

针对传统模糊聚类算法对初始聚类中心非常敏感以及对高光谱图像处理效果不佳的问题，为减少聚类数据的复杂度、降低聚类过程的计算成本以提升聚类性能，提出了一种基于多层二部图的高光谱模糊聚类算法。首先使用SuperPCA预处理方法对超像素分割得到的每个同质区域进行PCA来学习HSI数据不同区域的固有低维特征，从而获得高光谱数据的低维表示；其次，构造一个多层二部图矩阵来描述数据点和锚点之间的关系，降低了计算复杂度；最后，在模糊聚类中加入基于多层二部图的非负正则项来约束模糊隶属度矩阵的解空间。在Indian Pines和Pavia University数据集上进行的实验表明，所提算法能提高聚类效果与性能。

Fuzzy clustering based on hierarchical bipartite graph for large scale hyperspectral image

Traditional fuzzy clustering algorithm is very sensitive to the initial clustering center and has poor processing effect on hyperspectral image. To address this problem, this paper proposed a hyperspectral fuzzy clustering algorithm based on multilayer bipartit graph to reduce the complexity of clustering data and the computational cost of clustering process. Firstly, this paper used SuperPCA preprocessing method to perform PCA on each homogeneous region obtained by Super pixel segmentation to learn the inherent low-dimensional characteristics of different regions of HSI data, so as to obtain the low-dimensional representation of hyperspectral data. Then, this paper constructed a hierarchical bipartite graph matrix to describe the relationship between data points and anchor points, which reduced the computational complexity. At last, this paper constrained the solution space of the fuzzy membership matrix by adding the nonnegative regularization term based on hierarchical bipartite graph to fuzzy clustering. Experiments on Indian Pines and Pavia University datasets show that the proposed algorithm can improve the clustering effect and performance.