胶合板缺陷模糊神经网络检测算法

Algorithm of fuzzy-neural network in defect inspection of veneer wood

讨论了模糊逻辑和神经网络的工作原理,笔者在神经网络中引入模糊逻辑算法,将模糊逻辑处理不精确不完备信息的能力和神经网络的自适应自学习能力相结合,提出一种模糊神经网络算法,构建出模糊神经网络分类器,并且以胶合板缺陷检测为应用背景,对其分类的实时性、准确率等指标进行了验证,得到了分类精度93 33%,和训练次数5856次的良好性能.实验结果表明:在引入模糊逻辑算法后,基于模糊逻辑的神经网络分类器在模式分类精度和实时性等方面性能指标都得到了提高.

Theories of fuzzy logic and neural network are discussed in order to effectively deal with the inaccurate and incomplete information as well as the self-adjusting,self-learning and online self-adapting of neural network.An algorithm of fuzzy-neural network is carried out to solve the problems occurred during the process of pattern recognition,such as long adapting time of parameters and low classification accuracy.A classification machine based on the algorithm of fuzzy-neural network is set up as well.With the application of this method to the field of defect inspection of veneer wood,the accuracy of classification and rapidness of processing are validated,resulting in 93.33% accuracy of classification and 5,856 times of training.The experimental results show that high accuracy of pattern recognition and rapidness of processing have been obtained since applying the algorithm of the fuzzy logic to the classification based on fuzzy-neural network.