应用变分模态分解和随机森林特征选择算法的扬声器异常声分类

PDF(1548 KB)

振动与冲击 ›› 2022, Vol. 41 ›› Issue (20) : 277-283.

论文

周静雷，周智，崔琳

作者信息 +

Loudspeaker abnormal sound classification using variational modal decomposition and the random forest feature selection algorithm

ZHOU Jinglei， ZHOU Zhi， CUI Lin#br#

Author information +

文章历史 +

摘要

为了提高扬声器异常声分类的精度，提出了基于变分模态分解(variational mode decomposition，VMD)和随机森林特征选择算法的扬声器异常声分类方法。首先利用VMD分解采集到的扬声器声响应信号，之后对得到的一系列模态分量提取时域和频域特征；然后利用随机森林特征计算提取特征的重要性，通过递归特征消除算法提取出相关性较强的特征构造出最优特征子集；最后将最优特征子集输入至随机森林分类器中，实现扬声器异常声的分类识别。实验结果表明，该方法可以筛选出规模较小且识别度较高的低维特征数据集，同时具有更好的平均识别准确率，平均识别准确率为98.61%。
关键词：扬声器；异常声分类；变分模态分解；特征选择；随机森林

Abstract

In order to improve the accuracy of speaker abnormal sound classification, a speaker abnormal sound classification method based on variational mode decomposition (VMD) and random forest feature selection algorithm is proposed. Firstly, VMD is used to decompose the collected loudspeaker acoustic response signal, and then the time-domain and frequency-domain features of a series of modal components are extracted; Then, the importance of feature extraction is calculated by using random forest features, and the features with strong correlation are extracted by recursive feature elimination algorithm to construct the optimal feature subset; Finally, the optimal feature subset is input into the random forest classifier to realize the classification and recognition of speaker abnormal sound. The experimental results show that this method can screen out small and high recognition degree of low dimensional feature data sets, and has better average recognition accuracy, with an average recognition accuracy of 98.61%.
Key words: loudspeaker; abnormal sound classification; variational mode decomposition ;feature selection; random forest

导出引用

周静雷，周智，崔琳. 应用变分模态分解和随机森林特征选择算法的扬声器异常声分类[J]. 振动与冲击, 2022, 41(20): 277-283

ZHOU Jinglei， ZHOU Zhi， CUI Lin. Loudspeaker abnormal sound classification using variational modal decomposition and the random forest feature selection algorithm[J]. Journal of Vibration and Shock, 2022, 41(20): 277-283

参考文献

[1] Irrgang, Stefan, and Wolfgang Klippel. Fast and Sensitive End-of-Line Testing. Audio Engineering Society Convention 144. Audio Engineering Society, 2018: 9927-9937.
[2] Agerkvist F T, Torras-Rosell A, McWalter R. Improvements in Elimination of Loudspeaker Distortion in Acoustic Measurements. Audio Engineering Society Convention 138. Audio Engineering Society, 2015.
[3] Rumsey F . Loudspeaker design optimization and efficiency[J]. Journal of the Audio Engineering Society, 2018, 66(6):501-505.
[4] Thompson S , Pagliaro A , Celmer R , et al. Higher-order harmonic signature analysis for loudspeaker defect detection[J]. Journal of the Acoustical Society of America, 2003, 114(4):2400-2400.
[5] 周晓东,沈勇,薛政,等.窗函数在扬声器异常声客观检测中的影响[J].应用声学,2018,37(03):373-377.
ZHOU Xiao-dong , SHEN Yong, XUE Zheng , et al. The selection of window functions in loudspeaker Rub & Buzz detection. Journal of Applied Acoustics, 2018,37(03):373-377.
[6] 王鸿姗,周静雷,房乔楚.小波包样本熵的扬声器异常音特征提取方法[J].西安工程大学学报,2019,32(01):57-62.
WANG Hong-san,ZHOU Jing-lei,FANG Qiao-chu. Feature extraction method loudspeaker abnormal based on wavelet packet and sample entropy[J].Journal of Xi'an Polytechnic University, 2019,32(01):57-62.
[7] 韦峻峰,杨益,温周斌,等.一种扬声器异常音的时域特征检测方法[J].振动与冲击,2011,30(10):122-128.
WEI Jun-feng,YANG Yi,WAN Zhou-bin,et al. A detecting method for rub and buzz defects of a loudspeaker in time domain[J].Journal of Vibration and Shock,2011,30 (10) :122-128.
[8] 张雪芹.多分量非平稳信号的时频分析方法研究[D]. 山东大学, 2018.
[9] 张君,韩璞,董泽,等.基于小波变换的振动信号分析中能量泄漏的研究[J].中国电机工程学报,2004(10):240-245.
ZHANG Jun,HAN Pu,DONG Ze,et al. Energy Leakage research of Wavelet Transform application on Vibration Signal Analysis[J].Proceedingsof the CSEE,2004(10) : 240-245.
[10] Tao L, Zhijun L, Jiahong H, et al. A Comparative Study of Four Kinds of Adaptive Decomposition Algorithms and Their Applications[J]. Sensors, 2018, 18(7): 2120-2171.
[11] Dragomiretskiy K , Zosso D . Variational Mode Decomposition[J]. IEEE Transactions on Signal Processing, 2014, 62(3):531-544.
[12] Mohanty S , Gupta K K , Raju K S . Comparative study between VMD and EMD in bearing fault diagnosis[C]// IEEE International Conference on Industrial and Information Systems. IEEE, 2014.
[13] Liu C , Zhu L , Ni C . Chatter detection in milling process based on VMD and energy entropy[J]. Mechanical Systems & Signal Processing, 2018, 105(MAY15):169-182.
[14] Zeng W , Yuan J , Yuan C , et al. A new approach for the detection of abnormal heart sound signals using TQWT, VMD and neural networks[J]. Artificial Intelligence Review, 2020(1).
[15] 周静雷,颜婷.应用变分模态分解及能量熵的扬声器异常声分类[J].声学学报,2021,46(02):263-270.
ZHOU Jing-lei, YANG Ting. Loudspeaker abnormal sound classification using variational mode decomposition and energy entropy[J]. Acta Acustica,2021,46(02):263-270.
[16] Yu K , Guo X , Liu L , et al. Causality-based Feature Selection[J]. ACM Computing Surveys (CSUR), 2020.
[17] 薛瑞,赵荣珍.ReliefF与QPSO结合的故障特征选择算法[J].振动与冲击,2020,39(11):171-176+208.
XUE Rui,ZHAO Rong-zhen. The fault feature selection algorithm of combination of ＲeliefF and QPSO[J]. Journal of Vibration and Shock,2020,39(11):171-176+208.
[18] Ambusaidi, Mohammed A , Xiangjian, et al. Building an Intrusion Detection System Using a Filter-Based Feature Selection Algorithm.[J]. IEEE Transactions onuters, 2016.
[19] 许立武,李开成,罗奕,等.基于不完S变换与梯度提升树的电能质量复合扰动识别[J].电力系统保护与控制,2019,47(06):24-31.
XU Li-wu,LI Kai-cheng,LUO Yi,et al.Classification of complex power quality disturbances based on incomplete S-transform and gradient boosting decision tree[J]. Power System Protection and Control,2019,47(06):24-31.
[20] Huang J , Dong M , Lu S , et al. A hybrid model combining wavelet transform and recursive feature elimination for running state evaluation of heat-resistant steel using laser-induced breakdown spectroscopy[J]. Analyst, 2019.
[21] 李宏斌,徐楚林,温周斌.BP神经网络在扬声器异常音检测中的应用[J].声学技术,2014,33(06):522-525.
LI Hong-bin，XU Chu-lin，WEN Zhou-bin. The application of BP neural network in Loudspeaker’s Rub & Buzz Detection[J]. Technical Acoustics,2014,33(06):522-525.
[22] 郭庆,何劼恺,苏海涛,等.基于心理声学及支持向量机的扬声器异常音检测算法[J].东华大学学报(自然科学版),2020,46(02):275-281.
GUO Qing,HE Jie-kai,SU Hai-tao,et al. Speaker abnormal sound detection algorithm based on psychoacoustic model and support vector machine[J].Journal of Donghua University(Natural Science),2020,46(02):275-281.
[23] 武薇,申永军,杨绍普.基于排列熵理论的非线性系统特征提取研究[J].振动与冲击,2020,39(07):67-73.
Wu Wei,SHEN Yong-jun,YANG Sao-pu. Feature extraction of nonlinear system based on permutation entropy theory [J].Journal of Vibration and Shock,2020,39(07):67-73.
[24] 姚登举,杨静,詹晓娟.基于随机森林的特征选择算法[J].吉林大学学报(工学版),2014,44(01):137-141.
YAO Deng-ju,YANG Jing,ZHAN Xiao-juan. Feature selection algorithm based on random forest[J]. Journal of Jilin University(Engineering and Technology Edition), 2014,44(01):137-141.
[25] TEMME S, DOBOS V. Evaluation of audio test methods and measurements for end-of-line automotive loudspeaker quality control. Audio Engineering Society Convention 142, 2017: 1-10
[26] Bergstra J, Bengio Y. Random search for hyper-parameter optimization[J]. Journal of Machine Learning Research,2012,13(10):281-305
[27] 向阳辉,张干清,庞佑霞.结合SVM和改进证据理论的多信息融合故障诊断[J].振动与冲击,2015,34(13):71-77.
XIANG Yang-hui,ZHANG Gan-qing,PANG You-xia. Multi-information fusion fault diagnosis using SVM ＆ improved evidence theory[J].Journal of Vibration and Shock,2015,34(13):71-77