小波神经网络用于双模噪声的消噪

De-noising of Wavelet Neural Network for Bimodal Noise

小波神经网络利用了小波变换的良好的时域和频域的分析能力以及神经网络的自学习能力,具有良好的容错能力和逼近能力。针对双模斯噪声,提出基于小波神经网络的双模噪声背景下信号的消噪算法,介绍了双模噪声的3种简化模型,阐述了小波神经网络的基本概念以及基于此方法的消噪算法。将小波神经网络用于此3种双模噪声背景下信号的消噪。实验结果表明,该方法能有效地消除已知信号中的双模噪声。

WNN （Wavelet Neural Network）, with good time-and-frequency domain analysis capability of wavelet transform and self-learning ability, has strong fault tolerance and approach capability. For the non-Gaussian noise, the WNN-based signal de-noising algorithm under bimodal noisy background is proposed, the three simplified models of bimodal noise are described, the basic concepts of WNNexpounded, including the WNN-based denoising algorithm. WNN is applied to the signal denoising under this bimodal noisy background, and the experiment results indicate that this method could effectively eliminate the bimodal noise.