基于经验模态分解的水下目标回波提取

Empirical mode decomposition-based underwater target echo extraction

为在噪声中有效提取水下目标回波,提出一种基于经验模态分解的目标回波提取方法,分别对目标回波与噪声进行经验模态分解,得到二者各阶的固有模态信号分量;计算二者各阶的固有模态分量的能量,求取对应各阶的能量差,并与门限相比较,选取大于门限的固有模态分量重构目标回波信号.分析了经验模态分解阶数和选择门限等因素对重构信号精度的影响.仿真结果表明,该方法可精确提取目标回波.采用该方法提取信号的均方误差均低于自适应滤波方法,尤其在低信噪比时.

A signal extraction method based on empirical mode decomposition （EMD） was developed to extract underwater target echo from noise. The target echo and noise were decomposed respectively into their instinct mode signals （IMS） for each order by using EMD method. Each order of both target echo and noise IMS＇s energy was calculated, and energy difference for each order between target echo and noise was obtained. IMS components higher than threshold were chosen to reconstruct target echo. The influence of EMD orders and thresholds on precision of signal reconstruction was analyzed. Simulation results show that the method can extract target echo accurately. The mean square error of the proposed method is all less than that of adaptive filter method , especially in low signal noise ratio conditions.