文章在稀疏图约束NMF的基础上提出了一种基于加权残差协同图约束非负矩阵分解的高光谱解混算法（Weighted residual synergetic graph constraint NMF，WRGNMF）。在标准的NMF（Nonnegative Matrix Factorization）算法中，引入一个剩余加权机制，该策略是根据加权因子来处理残差中的值，为解混过程中每个原始像素与重构像素之间的重构误差提供适当的权值，让数据拟合更加精确，提高抗噪性能。同时，为了充分利用光谱图像的空间信息，用l₁范数来加强丰度矩阵的稀疏性，用图正则化来保持数据结构的亲和性。模拟和真实实验证实了该算法的有效性，提高解混精度的同时，对噪声更加鲁棒。

This paper proposes a hyperspectral unmixing algorithm based on the weighted residual synergetic graph constraint NMF(WRGNMF), building upon the foundation of sparse graph constraint NMF. In the standard NMF algorithm, a residual weighting mechanism is introduced. This strategy processes the values in the residual according to the weighting factor, providing appropriate weights for the reconstruction errors between each original pixel and the reconstructed pixel during the unmixing process, thereby enhancing data fitting accuracy and improving noise resistance. Simultaneously, to fully utilize the spatial information of spectral images, the l₁ norm is employed to enhance the sparsity of the abundance matrix, and graph regularization is used to maintain the affinity of the data structure. Both simulated and real experiments demonstrate the effectiveness of the proposed algorithm, which not only improves unmixing accuracy but also exhibits greater robustness to noise.