冲击回波信号的波形结构特征提取及分类研究

Research on Waveform Feature Extraction and Classification of Impact-Echo Signal

针对典型的非均质材料混凝土内部缺陷大小无损检测这一难题 ,利用人的听觉系统的生理学和心理学的研究成果 ,基于 EIH模型建立了进行时域波形结构特征提取的方法 ,并用结构自适应模糊聚类神经网络对提取的特征进行了分类试验 。

Concrete structure is a typical heterogeneous material. Flaw size detection of concrete structure is studied in this paper. Based on EIH model, a feature extraction method for impact-echo signal is proposed. An adaptive fuzzy clustering neural network classifier is used for classification of extracted features. The results of classification showed that this feature extraction method can resolve flaw size detection of concrete structure well.