摘要
鉴于现存研究中阈值函数存在尖点、收敛速度过快或过慢、阈值门限不能自适应不同分解系数等问题,提出了一种可调整的软阈值函数,并优化了基于动态惯性权重和正余弦振荡学习因子的粒子群算法,通过二者结合实现了信号的降噪。仿真结果表明:采用该方法优化的信号相较优化前信噪比提升了12.5%,粒子群算法的收敛速度提升了6倍,有效抑制了背景噪声的影响,提升了信号的可识别性。
Existing studies show issues with threshold functions,such as sharp points and unsuitable convergence rates,as well as thresholds unadaptable to different decomposition coefficients.To ad-dress those issues,a tunable soft threshold function was proposed and particle swarm algorithm was optimized based on dynamic inertia weights and sine-cosine oscillation learning factors.Combining the two approaches enables effective signal denoising.Simulation results demonstrate a 12.5%improve-ment in signal-to-noise ratio compared with the original signal,and the particle swarm algorithm's con-vergence speed is enhanced sixfold,effectively reducing background noise and improving signal recog-nizability.
作者
乔小瑞
张峻华
袁峰
QIAO Xiaorui;ZHANG Junhua;YUAN Feng(Naval Univ.of Engineering,Wuhan 430033,China;Unit No.92192,Ningbo 315000,China)
出处
《海军工程大学学报》
北大核心
2025年第4期73-78,共6页
Journal of Naval University of Engineering
关键词
海缆检测
信号去噪
小波变换
改进阈值函数
PSO
submarine cable inspection
signal denoising
wavelet transform
modified threshold function
PSO
