基于Morlet滤波器和粒子群算法的微弱信号检测

Weak Signal Detection Based on Morlet Wavelet Filtering and Particle Swarm Algorism

针对传统的微弱信号检测方法采用时域平均滤波方法或未优化的小波降噪法,具有测量次数多和信噪比低的缺点,提出了一种基于Morlet滤波器和粒子群算法的微弱信号检测方法;首先,给出了Morlet滤波器的降噪的具体方法,设计了实现最大信噪比的优化数学模型,然后以此数学模型作为适应度函数,定义了基于改进粒子群算法对半频带带宽d和中心频率fc进行优化的算法;为了进一步消除通带频段内的噪声,在传统阈值去噪的基础上,给出了采用最大似然估计法阈值降噪的具体方法;为了模拟实际环境,在实验中加入了均值为0方差为0.6的高斯噪声,实验结果表明:文中方法能有效地进行微弱信息检测,最大程度地还原为原始脉冲信号,且与其它方法相比,具有降噪效果好和可行性强的优点。

Aiming at the traditional weak signal detection method such as time domain filtering method and un optimized wavelet noise reduction method, having the defects of much numbers of measuring and the high ratio of signal to noise, a weak signal detection meth od based on Morlet wavelet filtering and particle swarm algorism was proposed. Firstly, the principle for Morlet wavelet noise deduction was described and the mathematical model for realizing the maximum signal to noise, then the improved particle swarm algorism was used to opti mizehalf band bandwidth d and center frequency f to deduct noise. In order to further eliminate noise in the passband frequency range, the specific method of maximum likelihood estimation method threshold noise reduction was given. The simulation experiment shows the method in this paper can realize weak signal detection and revert to the original pulse signal, and compared with other methods, with the ad vantages of godd noise reduction effect and strong feasibility.