摘要
针对人工监听识别飞机类型难度大的问题,提出了根据不同飞机发动机产生的不同噪声,通过特征提取,进而分类识别出飞机类型的一种方法。在梅尔倒谱系数(MFCC)算法特征提取的基础上,对提取的24维特征向量通过自编码器进行分类,对分类的准确率进行了仿真。实验结果表明,每一类声信号准确率均高于85%,且平均识别准确率为95.98%。针对单类别实际飞机声信号的分类准确率较其他类别准确率差的问题,提出了通过小波包分解-MFCC联合特征提取对自编码器进行优化。实验结果表明,每一类声信号准确率均高于90%,且平均准确率为97.74%。
Aiming at the difficulty of identifying aircraft types by artificial monitoring,a method of classifying and identifying aircraft types by extracting features according to the different noise produced by different aircraft engines is proposed.On the basis of feature extraction with MFCC algorithm,24 extracted feature vectors are classified by auto-encoder,and the classification accuracy is simulated.The experimental results show that the accuracy of each kind of acoustic signal is higher than 85%,and the average recognition accuracy was 95.98%.To solve the problem that the classification accuracy of single category of actual aircraft acoustic signals is lower than that of other categories,the auto-encoder is optimized through wavelet packet decomposition-MFCC combined feature extraction.The experimental results show that the accuracy of each kind of acoustic signals is higher than 90% and the average accuracy is 97.74%.
引文
[1] 李祥彬,李果,李学仁,等.飞机舱音记录器背景声的联合时频分析研究[J].应用声学,2009,28(1):53-58.
[2] 杨琳,王从庆,王芝刚,等.飞机舱音记录器非话语信号盲分离性能[J].南京航空航天大学学报,2010,42(2):185-190.
[3] 魏丹芳,李应.基于MFCC和加权动态特征组合的环境音分类[J].计算机与数字工程,2010,38(2):7-10.
[4] CHEN N,XIAO H D,WAN W.Audio Hash Function Based on Non-negative Matrix Factorisation of Mel-frequency Cepstral Coefficients[J].Information Security,IET,2011,5(1):19-25.
[5] ALAM M J,KENNY P,DOUGLAS O’ S.Low-variance Multitaper Mel-frequency Cepstral Coefficient Features for Speech and Speaker Recognition Systems[J].Cognitive Computation,2013,5(4):533-544.
[6] CHANG Y W,HSIEH C J,CHANG K W,et al.Training and Testing Low-degree Polynomial Data Mappings via Linear SVM[J].Journal of Machine Learning Research,2010,11(11):1471-1490.
[7] 戴礼荣,张仕良.深度语音信号与信息处理:研究进展与展望[J].数据采集与处理,2014,29(2):171-179.
[8] 邓俊锋,张晓龙.基于自动编码器组合的深度学习优化方法[J].计算机应用,2016,36(3):697-702.
[9] 崔江,唐军祥,龚春英,等.一种基于改进堆栈自动编码器的航空发电机旋转整流器故障特征提取方法[J].中国电机工程学报,2017,37(19):5696-5706.
[10] VINCENT P,LAROCHELLE H,LAJOIE I,et al.Stacked Denoising Autoencoders:Learning Useful Representations in a Deep Network with a Local Denoising Criterion[J].Journal of Machine Learning Research,2010,11(12):3371-3408.
[11] AREL I,ROSE D C,KARNOWSKI T P.Deep Machine Learning-A New Frontier in Artificial Intelligence Research[J].IEEE Computational Intelligence Magazine IEEE,2010,5(4):13-18.
[12] 王雅思.深度学习中的自编码器的表达能力研究[D].哈尔滨:哈尔滨工业大学,2014.
[13] 李祖贺,樊养余,王凤琴.YUV空间中基于稀疏自动编码器的无监督特征学习[J].电子与信息学报,2016,38(1):29-37.
[14] 贾棋,于美玉,樊鑫,等.基于曲率分级的形状编码及识别方法[J].计算机学报,2018,41(11):2453-2466.
[15] 邓俊锋,张晓龙.基于自动编码器组合的深度学习优化方法[J].计算机应用,2016,36(3):697-702.
[16] MILNER B,SHAO X.Prediction of Fundamental Frequency and Voicing From Mel-Frequency Cepstral Coefficients for Unconstrained Speech Reconstruction[J].IEEE Transactions on Audio,Speech & Language Processing,2006,15(1):24-33.
[17] WANG J C,WANG C Y,CHIN Y H,et al.Spectral-temporal Receptive Fields and MFCC Balanced Feature Extraction for Robust Speaker Recognition[J].Multimedia Tools & Applications,2016,76(3):1-14.
[18] PAVEZ E,SILVA J F.Analysis and Design of Wavelet-Packet Cepstral Coefficients for Automatic Speech Recognition[J].SpeEunication,2012,54(6):814-835.