利用模糊神经网络识别墙体材料的种类

Classification of Wall Board Materials by Fuzzy Neural Network Method

由超声脉冲法获得４ 种检测信息——声速、波幅、频率和波形，从中提取３ 个特征量：平均声速、声速的离差系数和波幅的离差系数，利用模糊分析方法和人工神经网络相结合对 Ｇ Ｒ Ｃ、加气混凝土、空心砖、实心砖、砌块 ５ 种墙体材料进行种类识别，形成一个智能化的、具有类似人脑功能的系统．经现场实测证明，判别的准确率达９５％ 以上。

Three characteristic features,such as average velocity of sound,deviation coefficient of sound velocity,and deviation coefficient of amplitude are extracted from four kinds of testing data which are sound velocity,amplitude,frequency and waveform in this paper.Fuzzy analysis and artificial neural network are combined to form an intelligent system with similar functions to human brain to classify GRC,aerated concrete,hollow brick,solid brick and concrete block.The accuracy of this system is up to 95% in field tests. [