人工神经网络在木材缺陷检测中的应用

Application of Artificial Neural Networks in the Testing of Wood Defects

采用射线作为检测手段,对木材进行无损检测。在无损检测信号处理和特征构造的基础上,运用特征参数建立了缺陷识别的数学模型,针对无损检测信号的特征,构造了人工神经网络。选用反向传播神经网络模型(BP网络),网络识别所需要的特征参数能够反映木材缺陷的全部特征。

In the paper, a nondestructive testing method of X-rays was adopted to detect the defects of wood. On the basis of processing signals and characteristics construction, the authors applied characteristic parameters to establish a mathematical modal of defects recognition and constructed artificial neural networks according to the characteristics of nondestructive signals. Back propagation networks was used to recognize all the characteristic parameters, which can reflect all the characteristics of wood defects.