基于模糊控制的小波包多阈值语音减噪新算法
|
摘要
在分析小波包传统阈值去噪算法的基础上,提出了一种基于模糊控制的小波包多阈值语音减噪新算法。该算法采用改进的多阈值选取方式来代替传统的阈值选择;应用一种新阈值函数对经小波包变换后的最底层频率系数进行量化处理以确保噪声尽可能地被滤除;模糊控制器可用于对信号中的幅值跳变以及边缘粗糙等问题进行优化与修正。综合以上3种方法即可自适应地进行语音增强处理。经实验结果验证,与传统阈值算法相比,该算法能够最大程度地还原纯语音信息,有效提高了语音去噪的准确度与信噪比。
Based on the analysis of wavelet packet traditional threshold denoising algorithm,a new algorithm of wavelet packet multi-threshold speech denoising based on fuzzy control is proposed. The algorithm adopted an improved multi threshold selection method instead of the traditional threshold selection; applied a new threshold function to quantize the bottom frequency coefficients after wavelet packet transform to ensure that the noise signal can be completely filtered out; fuzzy controller can be used to signal amplitude jump and edge roughness and other issues to optimize and correct. By combining these three methods,the speech enhancement process can be carried out adaptively. The experimental results show that compared with the traditional threshold algorithm,this algorithm restore the pure speech information to the greatest extent,effectively improve the accuracy of speech denoising and the signal-to-noise ratio.
Based on the analysis of wavelet packet traditional threshold denoising algorithm,a new algorithm of wavelet packet multi-threshold speech denoising based on fuzzy control is proposed. The algorithm adopted an improved multi threshold selection method instead of the traditional threshold selection; applied a new threshold function to quantize the bottom frequency coefficients after wavelet packet transform to ensure that the noise signal can be completely filtered out; fuzzy controller can be used to signal amplitude jump and edge roughness and other issues to optimize and correct. By combining these three methods,the speech enhancement process can be carried out adaptively. The experimental results show that compared with the traditional threshold algorithm,this algorithm restore the pure speech information to the greatest extent,effectively improve the accuracy of speech denoising and the signal-to-noise ratio.
引文
[1]彭辉,魏玮,陆建华.特定人孤立词的语音识别系统研究[J].控制工程,2011,18(3):397-400.Peng H, Wei W, Lu J H. Research on SpeakerDepended Isolated-Word Speech Recognition System[J].Control Engineering of China,2011,18(3):397-400.
[2] Donoho D L. De-noising by soft-thresholding[J]. IEEE Transactions on information theory,1995,41(3):613-627.
[3]陈晓曦,王延杰,刘恋.小波阈值去噪法的深入研究[J].激光与红外,2012,42(1):105-110.Chen X X,Wang Y J,Liu l. Deep study on wavelet threshold method for image noise removing[J]. Laser&Infrared,2012,42(1):105-110.
[4]王洪斌,王世豪,籍冰朔,等.基于改进多阈值小波包的去噪算法及应用[J].计量学报,2016,37(2):205-208.Wang H B,Wang S H,Ji B S,et al. An Improved Multiple Threshold Wavelet Packet De-noising Algorithm and Its Application[J]. Acta Metorlogica Sinaca,2016,37(2):205-208.
[5]郭晓霞,杨慧中.基于多阈值的小波包去噪[C]//中国自动化学会.第27届中国控制会议论文集.昆明:中国自动化学会,2009.
[6]吕振肃,马文.自适应小波阈值算法在心电信号去噪中的应用[J].数据采集与处理,2009,24(3):313-317.Lu Z S, Ma W. Application of adaptive wavelet threshold algorithm in ECG signal denoising[J].Journal of Data Acquisition and Processing,2009,24(3):313-317.
[7]秦学珍,潘俊涛,赵阳,等.基于模糊控制的小波消噪算法[J].电子世界,2016,(22):111-112.
[8]李庆春.新型PID模糊控制器的结构分析及应用研究[D].长沙:中南大学,2010.
[9]张花花.基于最优小波包基的信号增强算法研究及应用[J].电子测试,2010,(8):15-20.Zhang H H. Research and application of signal enhancement algorithm based on the wavelet packet[J].Electronic Test,2010,(8):15-20.
[10]谭文才,张秋菊.小波包多阈值去噪的一种改进方法[J].江南大学学报(自然科学版),2012,11(2):178-181.Tan W C,Zhang Q J. An Improved De-Noising Method by Wavelet Packet Multi-Threshold[J]. Journal of Jiangnan University(Natural Science Edition),2012,11(2):178-181.
[11] Bahoura M,Rouat J. A new approach for wavelet speech enhancement[C]//Proc European conf:on Speech communication and Technology. 2001:1 937-1 940.
[12]于向洋,罗志增.基于小波系数非线性连续函数衰减的脑电信号去噪[J].计量学报,2017,38(6):754-757.Yu X Y,Luo Z Z. EEG Signal Denoising Based on a Wavelet Nonlinear Continuous Function[J]. Acta Metorlogica Sinaca,2017,38(6):754-757.
[13]刘芳,邓志仁.基于自适应小波阈值和双边滤波器的去噪算法[J].系统仿真学报,2014,26(12):2 934-2 938.Liu F, Deng Z R. Denoising Algorithm Based on Adaptive Wavelet Threshold and Bilateral Filter[J].Journal of System Simulation,2014,26(12):2 934-2 938.
[14]杨阳,王贵君,杨永强.基于二叉树型分层的广义混合模糊系统推理规则数的缩减[J].控制理论与应用,2013,30(6):765-772.Yang Y,Wang G J,Yang Y Q. Reducing the number of inference rules for generalized hybrid fuzzy systems based on binary tree-type hierarchy[J]. Control Theory&Applications,2013,30(6):765-772.
[15]朱晓东,刘丹,李广.基于混合优化算法的模糊系统辨识[J].郑州大学学报(工学版),2015,36(4):10-14.Zhu X D,Liu D,Li G. Identification of Hierarchical Fuzzy System Based on Hybrid Optimization Algorithm[J]. Journal of Zhengzhou University(Engineering Science),2015,36(4):10-14.
[16]王宁,孟宪尧.两维最简模糊控制器结构分析[J].信息与控制,2008,37(1):34-39.Wang N,Meng X Y. Structure Analysis on the TwoDimensional Simplest Fuzzy Controller[J]. Information and Control,2008,37(1):34-39.
[17]宋知用. MATLAB在语音信号分析与合成中的应用[M].北京:北京航空航天大学出版社,2013:49-55.
[18]韦高梧.基于单信道的语音增强算法的研究与改进[D].广州:广东工业大学,2016.
[2] Donoho D L. De-noising by soft-thresholding[J]. IEEE Transactions on information theory,1995,41(3):613-627.
[3]陈晓曦,王延杰,刘恋.小波阈值去噪法的深入研究[J].激光与红外,2012,42(1):105-110.Chen X X,Wang Y J,Liu l. Deep study on wavelet threshold method for image noise removing[J]. Laser&Infrared,2012,42(1):105-110.
[4]王洪斌,王世豪,籍冰朔,等.基于改进多阈值小波包的去噪算法及应用[J].计量学报,2016,37(2):205-208.Wang H B,Wang S H,Ji B S,et al. An Improved Multiple Threshold Wavelet Packet De-noising Algorithm and Its Application[J]. Acta Metorlogica Sinaca,2016,37(2):205-208.
[5]郭晓霞,杨慧中.基于多阈值的小波包去噪[C]//中国自动化学会.第27届中国控制会议论文集.昆明:中国自动化学会,2009.
[6]吕振肃,马文.自适应小波阈值算法在心电信号去噪中的应用[J].数据采集与处理,2009,24(3):313-317.Lu Z S, Ma W. Application of adaptive wavelet threshold algorithm in ECG signal denoising[J].Journal of Data Acquisition and Processing,2009,24(3):313-317.
[7]秦学珍,潘俊涛,赵阳,等.基于模糊控制的小波消噪算法[J].电子世界,2016,(22):111-112.
[8]李庆春.新型PID模糊控制器的结构分析及应用研究[D].长沙:中南大学,2010.
[9]张花花.基于最优小波包基的信号增强算法研究及应用[J].电子测试,2010,(8):15-20.Zhang H H. Research and application of signal enhancement algorithm based on the wavelet packet[J].Electronic Test,2010,(8):15-20.
[10]谭文才,张秋菊.小波包多阈值去噪的一种改进方法[J].江南大学学报(自然科学版),2012,11(2):178-181.Tan W C,Zhang Q J. An Improved De-Noising Method by Wavelet Packet Multi-Threshold[J]. Journal of Jiangnan University(Natural Science Edition),2012,11(2):178-181.
[11] Bahoura M,Rouat J. A new approach for wavelet speech enhancement[C]//Proc European conf:on Speech communication and Technology. 2001:1 937-1 940.
[12]于向洋,罗志增.基于小波系数非线性连续函数衰减的脑电信号去噪[J].计量学报,2017,38(6):754-757.Yu X Y,Luo Z Z. EEG Signal Denoising Based on a Wavelet Nonlinear Continuous Function[J]. Acta Metorlogica Sinaca,2017,38(6):754-757.
[13]刘芳,邓志仁.基于自适应小波阈值和双边滤波器的去噪算法[J].系统仿真学报,2014,26(12):2 934-2 938.Liu F, Deng Z R. Denoising Algorithm Based on Adaptive Wavelet Threshold and Bilateral Filter[J].Journal of System Simulation,2014,26(12):2 934-2 938.
[14]杨阳,王贵君,杨永强.基于二叉树型分层的广义混合模糊系统推理规则数的缩减[J].控制理论与应用,2013,30(6):765-772.Yang Y,Wang G J,Yang Y Q. Reducing the number of inference rules for generalized hybrid fuzzy systems based on binary tree-type hierarchy[J]. Control Theory&Applications,2013,30(6):765-772.
[15]朱晓东,刘丹,李广.基于混合优化算法的模糊系统辨识[J].郑州大学学报(工学版),2015,36(4):10-14.Zhu X D,Liu D,Li G. Identification of Hierarchical Fuzzy System Based on Hybrid Optimization Algorithm[J]. Journal of Zhengzhou University(Engineering Science),2015,36(4):10-14.
[16]王宁,孟宪尧.两维最简模糊控制器结构分析[J].信息与控制,2008,37(1):34-39.Wang N,Meng X Y. Structure Analysis on the TwoDimensional Simplest Fuzzy Controller[J]. Information and Control,2008,37(1):34-39.
[17]宋知用. MATLAB在语音信号分析与合成中的应用[M].北京:北京航空航天大学出版社,2013:49-55.
[18]韦高梧.基于单信道的语音增强算法的研究与改进[D].广州:广东工业大学,2016.