改进小波阈值-形态学融合算法在振动信号误差抑制中的应用

Application of Improved Wavelet Threshold and Morphological Integrated Algorithm in Vibration Signal Error Suppression

针对机械装备振动信号采集中含有脉冲噪声及高斯噪声等多种复杂污染源的信号误差抑制去噪,设计了小波阈值和形态学融合算法。首先,在高斯平滑思想上引入改进的小波阈值函数,能够调节阈值函数向小波分解系数真值的收敛趋势达到有效去除高斯噪声及避免稳态偏差的目的,对比几种常用的函数算法,提出的阈值函数取得了在信噪比提升及均方根误差抑制度量上的更优的效果。其次,对比计算了多种形态学算子选择较优的算法用于去除脉冲噪声。最后,将融合算法应用在机械臂振动信号的去噪实验中。结果表明,算法能够有效去除噪声并保留信号特征。

In this paper,an integrated algorithm combining wavelet threshold denoising method with morphological denoising method was designed to suppress the signal error and denoise the complex pollution sources including impulse noise and Gaussian noise in the vibration signal of mechanical equipment.Firstly,the improved wavelet threshold function was introduced based Gaussian smoothing method,which could adjust the convergence trend of the threshold function to the true value of the wavelet transforms coefficient to effectively remove the Gaussian noise and avoid the steady-state deviation,the threshold function proposed had a better effect on the measurement of SNR and RMSE compared with several commonly used algorithms.Secondly,the best algorithm was selected from several morphological operators for pulse noise removal.Finally,the integrated algorithm was applied to the denoising experiment of the vibration signal of the manipulator.The results showed that the algorithm could effectively remove the noise and retain the signal characteristics.