小波包分析在木材超声检测中的应用

Application of Wavelet Packet Analysis in Wood Ultrasonic Testing

小波包分解能够准确地把超声检测信号划分到不同的频带范围内,实现了不同频带范围内能量变化量的提取。试验说明缺陷试件的信号能量变化量在一些特定的频带内显示出显著的不同,为下一步应用神经网络等模式识别方法辨别试件缺陷类型提供了比较理想的特征向量。

Wavelet packet can accurately divide ultrasonic signals into different frequencies and realize abstraction of energy variation.The results show that energy variation of signals of defect samples are obviously different in some certain frequencies,which provide perfect eigenvalues for Artificial Neural Network to identify the types of defects.