一种优化的语音信号处理方法

An optimized voice signal processing method

为了解决传统利用压缩感知采集语音信号的方法中所用的测量矩阵难以硬件实现,且所用的重构方法过于繁琐导致重构时间过长的问题,采用0-1随机矩阵对语音信号进行测量,用分段正交匹配追踪算法对语音信号进行重构.实验结果表明,优化后的方法可以很好地解决传统压缩感知中硬件难以实现和重构算法繁琐的问题.该方法中所用的0-1随机矩阵便于硬件实现,所用的分段正交匹配追踪(StOMP)算法能够加速信号的重构.

To solve the problem that the measurement matrix used in the traditional method of using compressed sensing to collect voice signals is difficult to implement in hardware,and the reconstruction method used is too cumbersome,resulting in a long reconstruction time,a 0-1 random matrix is used to measure the speech signal,and stagewise orthogonal matching pursuit algorithm is used to reconstruct the speech signal.The experimental results show that the optimized method can well solve the problem of difficult hardware implementation and cumbersome reconstruction algorithm in traditional compressed sensing.The 0-1 random matrix used in this method is convenient for hardware implementation,and stagewise orthogonal matching pursuit algorithm used can speed up signal reconstruction.