希尔伯特-黄变换在微弱被动瞬态鱼声信号中的检测

Research on the detection of weak transient passive fish acoustic signals based on Hilbert–Huang Transform

为了在海洋环境中准确获取鱼类在摄食、产卵和争斗过程中发出的短促、瞬时的信号,采用希尔伯特-黄变换算法实现微弱被动瞬态鱼声信号的检测。该算法首先将瞬态鱼声信号实现固有模态信号的分解,其次将经验值高阶阶数的固有模态信号求和重构即可实现信号能量幅度检测;在固有模态信号的分解的基础上,计算求和重构信号的希尔伯特能量谱,即可实现瞬时能量密度级的检测。通过不同信噪比和检测器的比较研究,结果表明基于希尔伯特-黄变换算法的两种检测方法能有效提高微弱被动瞬态鱼声信号的检测性能。

To acquire transient passive fish acoustic signals during feeding and spawning, Hilbert–Huang Transform is introduced to detect weak signals. Firstly, Hilbert-Huang Transform analyzes the transient signals in the time-frequency domain by empirical mode decomposition. Energy detection is then realized by Hilbert–Huang Transform reconstruction from empirical high-level orders. In addition, Hilbert spectral analysis is proposed to detect transient signals and compared with a traditional energy method under a different signal-to-noise ratio（SNR）. Finally, the feasibility and validity of the method are verified by processing experimental data.