基于稀疏约束非负矩阵分解的水下线谱增强方法

Underwater spectral line enhancement based on non-negative matrix factorization with sparseness constraint

水下目标线谱增强是被动声纳目标探测的关键问题之一。传统的线谱信号处理方法集中于时域和频域处理,本文提出了利用非负矩阵分解,在时频联合域内进行线谱信号增强的处理方法,以目标线谱信号的时频矩阵作为非负矩阵分解的输入,通过基矩阵提取线谱信号的频谱模式。根据线谱信号在频域的稀疏性质,对基矩阵进行稀疏性约束,利用权重稀疏扫描的方式讨论基矩阵的稀疏度和频率估计精度随权重系数的变化关系,确定稀疏约束项的有效权重系数区间。仿真结果显示,稀疏约束项在低信噪比条件下表现出优越的线谱增强能力,最低信噪比可达-30 dB。海试数据结果表明,此方法可以有效地提高对线谱信号的提取能力。

Spectral line enhancement of underwater objects is one of the critical issues for passive sonar systems.Conventional approaches for processing spectral lines have focused on either time-domain or frequency-domain methods.The non-negative matrix factorization is proposed to process the underwater spectral lines in the joint time-frequency domain.The time-frequency matrix of the spectral lines is utilized as the input of non-negative matrix factorization and the frequency-mode of the spectral lines is extracted by the basis matrix.Based on the sparsity of spectral lines in the frequency domain,the sparseness constraint is utilized to constrain the basis matrix.The correlation between sparsity and frequency estimation accuracy is examined with weight coefficients scanning,and an effective weight coefficient interval for the sparseness term is determined.Simulation results show that the sparseness term exhibits superior spectral line enhancement ability under low SNR conditions,with the lowest SNR reaching about-30 dB.Sea experimental results demonstrate that the algorithm significantly enhances the capacity for the extraction of spectral lines.