水声脉冲信号检测是一水声探测问题，对于不同信号形式与参数，频谱熵检测法为一种有效且可行的检测方法，但对于频域能量聚集性弱的信号形式检测性能欠佳。理论分析并利用循环谱检测法特点，结合熵检测算法，提出频谱熵与循环谱联合检测方法，提高了对频域能量聚集性弱的信号形式检测性能。并对检测过程中检测曲线的滤波需求，分析高斯滤波与双边滤波原理，提出改进双边高斯滤波方法，达到对检测曲线较好平滑且保留脉冲信号突变特征的理想滤波效果。结果表明所提检测及滤波方法相比于常规信号检测方法检测率均有提高。
Underwater acoustic pulse signal detection is crucial in sonar applications. Spectral entropy detection effectively handles various signal types and parameters but performs poorly for signals with weak spectral energy concentration. This study theoretically analyzes and utilizes cyclic spectrum detection characteristics alongside entropy detection algorithms to propose a joint spectral entropy and cyclic spectrum detection method. This approach enhances detection performance for signals with weak frequency domain energy aggregation. Additionally, to address the need for filtering detection curves, Gaussian and bilateral filtering principles are examined, and an improved bilateral Gaussian filter is introduced. This filter smoothly processes detection curves while preserving the abrupt features of pulse signals. Simulation results demonstrate that the proposed detection and filtering methods effectively detect seven common sonar signals, achieving higher detection rates compared to conventional methods.
2025,47(18): 142-148 收稿日期：2024-12-5
DOI：10.3404/j.issn.1672-7649.2025.18.023
分类号：U666.7；TN911.7
作者简介：查淞元（1999 – ），男，硕士研究生，研究方向为水声信号处理
参考文献：
[1] 田坦. 声呐技术. 第2版[M]. 哈尔滨工程大学出版社, 2010.
[2] 王晓燕, 方世良, 朱志峰. 非合作水声脉冲信号的熵检测方法[C]//中国声学学会. 泛在信息社会中的声学——中国声学学会2010年全国会员代表大会暨学术会议论文集. 东南大学水声信号处理教育部重点实验室, 2010.
[3] 赵安琪. 未知参量水声脉冲信号检测技术研究[D]. 哈尔滨: 哈尔滨工程大学, 2018.
[4] ROBERT M, Entropy and information theory[J]. Springer-Verlag, 1990.
[5] ZHANG Y, ZHANG Q, WU S. Entropy-based robust spectrum sensing in cognitive radio[J]. IET Communications, 2010(4): 4-8.
[6] 申达. 认知无线电频谱检测技术研究[D]. 上海: 上海交通大学, 2010.
[7] 李剑飞. 基于稀疏理论的高时相高空间分辨率遥感图像模拟[D]. 葫芦岛: 辽宁工程技术大学, 2016.
