作者：温益凯, 陈乐, 富雅琼

作者单位：中国计量大学机电工程学院，杭州 浙江 310018

关键词：蜂鸣声识别;小波变换;降噪;软阈值;尺度系数;时频分析

摘要：

针对工厂复杂环境中电子产品的蜂鸣声信号信噪比低所造成的难以识别问题，采用小波变换对蜂鸣声信号进行降噪处理。为获得良好的降噪效果，对不同小波降噪方法以及不同小波参数选取的降噪效果进行对比分析，得出采取db1小波基以sqtwolog为阈值标准的软阈值法去噪效果良好，但是软阈值法随着小波分解层数的递增，容易造成信号失真。通过对蜂鸣声信号小波包分解节点的分析，最终验证小波包尺度系数比对法去噪效果最好且不会引起信号失真，去噪后信号的RMSE为0.0094，SNR为14.7513 dB，并且基于处理后的信号利用小波的时频分析特点可以快速识别蜂鸣声信号。

A fast recognition method of buzzer based on wavelet denoising
WEN Yikai, CHEN Le, FU Yaqiong
College of Mechanical and Electrical Engineering, China Jiliang University, Hangzhou 310018, China
Abstract: Aiming at the difficult identification problem caused by the low signal-to-noise ratio of the buzzer signal of electronic products in the complex environment of the factory, wavelet transform was used to reduce the noise of the buzzer signal. In order to obtain a good denoising effect, the denoising effect of different wavelet denoising methods and different wavelet parameters was compared and analyzed, and it is concluded that the soft threshold method with db1 wavelet base and sqtwolog as the threshold standard has good denoising effects, but with the increase of wavelet decomposition level, the soft threshold method was easy to cause signal distortion. Through the analysis of the wavelet packet decomposition node of the buzzer signal, it is finally verified that the wavelet packet scale coefficient comparison method has the best denoising effect and does not cause signal distortion. The RMSE of the denoised signal is 0.0094, and the SNR is 14.7513 dB. And based on the processed signal, the time-frequency analysis characteristics of wavelet can be used to quickly identify the buzzer signal.
Keywords: buzzer recognition;wavelet transform;noise reduction;soft threshold;scale factor;time-frequency analysis
2022, 48(4):12-17,34  收稿日期: 2021-01-18;收到修改稿日期: 2021-03-09
基金项目:
作者简介: 温益凯（1996-），男，温州苍南县人，硕士研究生，专业方向为在线检测与控制系统
参考文献
[1] 胡柏青, 魏峥, 王伯雄, 等. 强噪条件下基于小波降噪的陀螺仪声信号处理方法[J]. 传感技术学报, 2008(6): 109-111
[2] 罗海涛. 用小波消除语音噪声[J]. 福建电脑, 2017, 33(8): 119-120
[3] 赵景波, 唐勇伟, 张磊. 基于改进小波变换的故障电弧检测方法的研究[J]. 电机与控制学报, 2016, 20(2): 90-97
[4] 周迪, 刘昌明, 王志刚, 等. 小波变换在混凝土冲击回波检测中的应用[J]. 中国测试, 2019, 45(4): 135-140
[5] 向瑾, 翟成瑞, 杨卫, 等. 基于小波变换的音频信号去噪[J]. 微计算机信息, 2007, 23(35): 85-86
[6] EI-SHEIMY N, NASSAR S, NOURELDIN A. Wavelet de-noising for IMU alignment[J]. IEEE Aerospace & Electronic System Magazine, 2004, 19(10): 32-39
[7] 崔桂梅, 姚艳清, 张勇. 小波域BM3D滤波算法高炉回旋区温度场测量[J]. 中国测试, 2019, 45(11): 9-13, 20
[8] 刘书俊, 李生林, 蒋明, 等. 一种基于加权平均的改进型小波阈值降噪算法[J]. 化工自动化及仪表, 2017, 44(3): 239-242, 318
[9] 许亚男, 王旭升. 基于小波包阈值方法的齿轮箱振动信号降噪处理[J]. 电脑知识与技术, 2016, 12(23): 238-240
[10] 章浙涛, 朱建军, 匡翠林, 等. 小波包多阈值去噪法及其在形变分析中的应用[J]. 测绘学报, 2014, 43(1): 13-20
[11] 苏建芳, 吴钦木. 基于小波包分析的电机滚动轴承故障诊断[J]. 测控技术, 2019, 38(4): 64-67
[12] 朱宁, 徐常新, 符影杰. 超声波测距系统中的小波去噪方法研究[J]. 仪表技术与传感器, 2021(1): 102-106, 122
[13] 邸亚洲, 李富荣, 于建立, 等. 小波分析在飞参数据降噪中的应用[J]. 计算机仿真, 2010, 27(10): 1-4