集成滤波器在谱图信号除噪中的应用

Integration Filter Applied to the Noise Removing From Spectral Signal

本文提出了一种由神经元网络与线性自适应滤波器组成的集成滤波器的实现方法，用于谱图信号除噪处理。集成滤波器采用神经元网络描述谱图信号的非线性特性，采用自适应滤波算法提高除噪能力。实验证明集成滤波器能在较小的信号失真下具备良好的除噪能力。

In this paper, a novel methodology of integration filter made up of artificial neural network (ANN) and linear adaptive filter (LAF) was proposed to remove noise from spectral signal(SS). In the integration filter, the nonlinear property of SS is represented by ANN,also adaptive algorithm is used to improve the ability of noise removing. It came to the conclusion that the integration filter could have a powerful capability of noise removing under the condition of less signal distortion according to the experiment.