新的联合三角形法和经验模态分解的水下目标回波检测

水下目标回波信号的自动检测和识别是水声装备自动化的关键任务之一。大量的混响干扰,一个脉冲周期内有多个目标或没有目标,以及目标边界的模糊性和位置的随机性,是目标检测的难点。对回波信号进行经验模态分解,目标信号的绝大部分信息将集中在某个固有模态函数上,且产生明显的边界,而各种噪声和干扰成分被分解到其他固有模态函数上。基于此,提出在固有模态函数包络上用三角形法定位目标及双阈值检测精确的目标边界的算法;提出目标回波的固有模态函数瞬时幅度方差谱特征。海上实测回波数据的实验,从检测目标个数和识别率两方面来验证算法的性能,并与其它方法进行对比。实验表明该算法能任意定位和准确检测回波中的目标,且不受脉冲噪声的影响,检测效率高。

Detection of underwater target echo with triangle method associated with empirical mode decomposition

It is one of crucial tasks in underwater acoustic equipment development to detect and recognize underwater target echo automatically.Difficulty of detection of target echo lies in much reverberation,no target or more than one target existing in an impulse period,illegibility of targets' edges and randomicity of targets' location.Most of information about target signal is concentrated in intrinsic mode function on some scale after empirical mode decomposition being done for target echo.It will produce obviou...