进而构造单源领域特征矩阵，特征分解取最大特征值对应的特征向量，通过聚类

Further features art single-source configuration matrix, eigenvector decomposition feature takes the maximum eigenvalue corresponding to the estimated mixing matrix clustering completion signal and the number of frequency hopping;

Then construct the characteristic matrix of single source field, the feature decomposition takes the characteristic vector corresponding to the maximum feature value, and the number estimation of the mixed matrix and frequency frequency signal is completed by clustering.

Then, a single source domain eigenmatrix is constructed, and the eigenvectors corresponding to the maximum eigenvalues are extracted by eigendecomposition;<br>