在强混响背景下，使用传统的预白化处理、时频分析以及子空间分析等方法对动目标检测效果不佳，针对这一问题，本文利用近年来新引入的低秩稀疏矩阵分解理论来提高强混响背景下的动目标检测能力，采用多帧数据联合的鲁棒PCA处理算法，结合混响数据的声学特征将声学检测问题转化为图像分解问题，并通过对比PCA算法处理结果，给出算法的性能比较；与此同时，本文结合目标运动连续性和稀疏杂点随机性的特征差异，提出一种定位窗滤波方法，进一步滤除稀疏杂点，净化主动声呐显示图像，提高主动声呐动目标检测性能。仿真及试验数据处理结果说明，在阵元端信混比–5 dB情况下，算法仍然可以对目标准确定位，滤除稀疏杂点，且在时频域上效果更佳，显著提高了主动声呐动目标检测能力。
In the context of strong reverberation, traditional methods such as pre-whitening, time-frequency analysis and subspace analysis are not effective in detecting moving targets. To address this issue, this paper uses the newly introduced low-rank sparse matrix decomposition theory in recent years to improve the detection ability of moving targets in the context of robbery. A robust PCA processing algorithm combining multiple frames of data is adopted. By combining the acoustic characteristic of reverberation data, the acoustic detection problem is transformed into an image decomposition problem, and the performance comparison of the PCA algorithm is given by comparing the processing results. At the same time, this article proposes a localization window filtering method that combines the feature differences of target motion continuity and sparse clutter randomness, further filtering out sparse clutter, purifying active sonar display images, and improving the performance of active sonar moving target detection. The simulation and experimental data processing results show that under the signal-to-noise ratio of –5 dB at the array end, the algorithm can still accurately locate the target, filter out sparse clutter, and achieve better results in the time-frequency domain, significantly improving the active sonar moving target detection ability.
2024,46(20): 153-158 收稿日期：2023-12-15
DOI：10.3404/j.issn.1672-7649.2024.20.028
分类号：U666.7
作者简介：马怀逸(1997-),男,硕士研究生,研究方向为水声信号处理
参考文献：
[1] 陈云飞, 李佳娟, 王振山, 等. 水中目标回波亮点统计特征研究[J]. 物理学报, 2013, 62(8): 284-294.
[2] 高博, 杨士莪, 朴胜春, 等. 界面混响的耦合简正波理论研究[J]. 声学学报, 2013, 38(5): 517-522.
[3] 郭富强, 杨益新, 孙超, 等. 浅海低频本地海底混响的波导不变性结构及抑制方法研究[J]. 声学学报, 2009, 34(6): 506-514.
[4] TOH K-C, YUN S. An accelerated proximal gradient algorithm for nuclear norm regularized least squares problems[J]. Pacific Journal of Optimization, 2010, 6(3): 615-640.
[5] CHEN M. Fast convex optimization algorithms for exact recovery of a corrupted low-rank matrix[C]//Coordinated Science Laboratory U O I A U-C, 2009.
[6] LIN Z, CHEN M, MA Y. The augmented lagrange multiplier method for exact recovery of corrupted low-rank matrices. mathematical programming[J]. Eprint Arxiv, 2010, 9.
[7] ZHOU T, TAO D. GoDec: randomized lowrank & sparse matrix decomposition in noisy case [C]//Proceedings of the 28th International Conference on Machine Learning, ICML 2011. 2011.
[8] LI W, SUBRAHMANYA N, XU F. Online subspace and sparse filtering for target tracking in reverberant environment[C]//IEEE 7th Sensor Array and Multichannel Signal Processing Workshop, Hoboken, NJ, USA, 2012.
[9] GE F X, CHEN Y, LI W. Target detecton and tracking via structured convex optimization[C]// Icassp IEEE International Conference on Acoustics, IEEE, 2017.
[10] 刘建设, 殷敬伟, 朱广平, 等. 冰下运动目标主动探测技术研究[J]. 应用声学, 2019, 38(4): 562-568.
[11] 朱广平, 宋泽林, 殷敬伟, 等. 混响背景下低秩矩阵恢复的目标亮点特征提取[J]. 声学学报, 2019, 44(4): 471-479.
[12] 吴耀文, 邢传玺, 岳露露, 等. 基于鲁棒主成分分析的低频水声信号降噪方法[J]. 云南民族大学学报(自然科学版), 2020, 29(1): 70-77.
[13] 陈勇. 不确定海洋环境中凸优化稳健波束形成与方位估计研究[D]. 长沙: 国防科技大学, 2018.
[14] 贺玉梁. 运动小平台主动声呐目标回波信号检测技术研究[D]. 哈尔滨: 哈尔滨工程大学, 2020.
