Research On Vehicle Noise Suppression Based On Wavelet Transform And Singular Value Decomposition

In the field of audio signal processing,noise reduction has always been a key research topic.Obtaining a clean target signal from a signal affected by noise is of great significance for subsequent audio recognition or signal detection accuracy improvement.In the process of transmitting and receiving signals,smart vehicle navigation is often interfered by complex vehicle noise,which makes the received signal waveforms appear distorted and distorted,which in turn causes the navigation wireless digital receiving system terminals to fail to recognize useful signals,resulting in mixed signals.On the coding information.In this paper,a noise reduction method combining wavelet threshold processing and singular value decomposition processing is used to study the noise suppression processing of one-dimensional audio signals received by intelligent navigation contaminated by vehicle noise.The main research work is as follows:1)Firstly,the theoretical basis of wavelet transform is described.Based on the characteristics of noise processing by wavelet transform,a wavelet threshold denoising method is proposed,and then the characteristics of different thresholds and threshold functions are explained in detail.The threshold function guarantees the smoothness of the waveform after signal noise reduction.By comparing the main characteristics and commonly used indicators of wavelet basis,four commonly used wavelet basis functions are obtained for the decomposition and reconstruction experiment of multi-resolution analysis,and it is determined that the Sym6 wavelet basis function has the best noise reduction effect.At the same time,the range of the number of layers is determined by the wavelet adaptive layering algorithm,and a comparative experiment is made within the range.2)Secondly,Aiming at the problem that the singular value decomposition algorithm takes too long to process,the idea of segmentation and re-decomposition is proposed to reduce noise and improve at the same time.The operation efficiency is improved,and the running program will not be redundant.According to the feature space of the noise signal and the useful signal are not related to each other,the singular value decomposition method to determine the threshold is discussed,and it is concluded that the effective rank order is determined by the singular value decomposition variance threshold method,which effectively extracts the usefulness the feature space of the signal.3)Finally,the wavelet transform and singular value decomposition are combined effectively,the WT-SVD joint model algorithm is proposed.After experimental simulation,it passes objective quantitative indicators: signal-to-noise ratio,root mean square error,and subjective auditory effect: The comparative analysis of mean opinion scores can draw a conclusion: in the environment of vehicle noise interference,for the noise suppression effect of the signal received by the intelligent navigation system,the joint algorithm has good adaptability and good noise suppression compared with the single wavelet transform noise reduction algorithm.performance.The research in this paper further strengthens the anti-interference ability of the received signal of the vehicle’s intelligent navigation system against vehicle noise,deepens and enriches the research of audio signal on noise suppression,and lays a certain foundation for the subsequent audio signal recognition work.