模糊神经网络的极点配置自校正滤波器

A Self-Corrected Filter Based on Fuzzy Neural Network and Pole Allocation

将模糊Hopfield神经网络应用于间接自校正滤波器设计 ,提出用一阶Taylor逼近神经元输出作为参数估计的方法 .将模糊Hopfield神经网络应用于自校正滤波中 ,并将该滤波器设计方法应用于图像去噪 ,得到了较满意的滤波结果 .表明使用模糊Hopfield网络极点配制自校正控制方法的滤波器设计有较好的实用性 .

The fuzzy Hopfield neural network is applied to the design of a indirectly self corrected filter, and the first order Taylor approximate neuron output is taken as the method of parameter estimation. Moreover, the fuzzy Hopfield neural network is applied to self corrected filtration; and the design method of this filter is applied to image denoising, which results in a fairly satisfactory filtration. The results indicate that it is fairly practical to design a filter by using self corrected control method based on fuzzy Hopfield network and pole allocation.