基于模糊距离的RBF神经网络板形模式识别

针对BP网络在板形缺陷模式识别的应用中,由于网络本身存在收敛速度慢,学习时间长,容易陷入局部极小等结构性能不佳的问题,从而影响板形识别精度;文中提出了基于模糊距离的RBF网络板形缺陷模式识别法.RBF网络是一种局部逼近的神经网络,它在逼近能力、分类能力和学习速度等方面都优于BP网络.为了使板宽变化时网络结构不变,将模糊距离与RBF网络相结合进行模式识别,仿真实验结果表明,该方法与BP网络方法相比,网络收敛速度较快,而且识别精度较高.

Using RBF Network Based on Fuzzy Distance to Flatness Pattern Recognition

This paper puts forward using RBF network based on fuzzy distance to flatness pattern recognition to improve recognition precision.RBF network is a kind of part approach to neural network.It is more capable of approach and classification and is higher speed of learning than BP network.Simulation results of RBF network method based on fuzzy distance indicate that is able to fix network structure as change range of armor plate.that the method is higher speed convergence and better recognition precision than BP network.