基于Teager能量算子的改进阈值函数的去噪算法研究 postprint

Abstract: For the problems such as the aliasing phenomenon of signal and noise wavelet packet coefficients, threshold function discontinuity at threshold, and constant deviation between original coefficient and estimation coefficient in traditional threshold denoising, this paper proposed an improved wavelet packet threshold denoising algorithm. The algorithm first calculated the Teager energy operator for the wavelet packet coefficients, making the difference between voice and noise coefficients larger, which is favorable for threshold selection. Then, for the problems of pseudo Gibbs effect and constant deviation, this paper proposed an improved threshold function, which not only overcomes the problem of discontinuity and constant deviation of commonly used functions, but also has superior mathematical characteristics. The experimental results show that the improved algorithm has higher signal-to-noise ratio and reduces the mean square error, which demonstrates the algorithm eliminates the noise to the greatest extent and avoids the signal distortion, and has higher practical value.