基于多分辨率奇异熵的混凝土缺陷检测

发布时间：2018-03-25 23:56

本文选题：小波变换　切入点：奇异谱分析　出处：《系统仿真学报》2017年04期

【摘要】：针对超声波检测信号存在衰减大、指向性差、传播路径复杂,以及构成复杂的特点,提出一种基于多分辨率奇异熵的混凝土缺陷智能无损检测算法。该算法利用小波算法,将超声波信号分解为多个尺度下的高、低频分量,针对各个分量进行奇异谱分解,同时利用信息熵理论,计算奇异熵作为缺陷检测的特征值,利用GA-SVM算法对奇异熵进行训练,从而达到辨识混凝土缺陷的目的。仿真实验结果表明表明采用该方法能够有效的提升混凝土缺陷辨识的精确度,提供了一种新的辨识混凝土缺陷的有效途径。
[Abstract]:Aiming at the characteristics of large attenuation, poor directivity, complex propagation path and complex structure of ultrasonic detection signal, an intelligent nondestructive detection algorithm for concrete defects based on multi-resolution singular entropy is proposed. The ultrasonic signal is decomposed into high and low frequency components of multiple scales, and the singular spectrum is decomposed for each component. At the same time, the singular entropy is calculated as the eigenvalue of defect detection by using the information entropy theory, and the singular entropy is trained by GA-SVM algorithm. The simulation results show that this method can effectively improve the accuracy of concrete defect identification and provide a new effective way to identify concrete defects.
【作者单位】：中南大学防灾科学与安全技术研究所;
【基金】：国家自然科学基金(51208525)
【分类号】：TB559;TU755.7

【相似文献】