基于多分辨率奇异熵的混凝土缺陷检测  

Forestry Fire Spatial Diffusion Model Based on Integration of Multi-Agent Algorithm with Cellular Automata

作　　者：赵迪[1] 徐志胜[1] Zhao Di Xu Zhisheng(Institute of Disaster Prevention Science and Safety Technology, Central South University, Changsha 410075, China)

机构地区：[1]中南大学防灾科学与安全技术研究所,长沙410075

出　　处：《系统仿真学报》2017年第4期761-766,共6页Journal of System Simulation

基　　金：国家自然科学基金(51208525)

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。In view of the existing ultrasonic detection signal attenuation, directional difference, complex propagation paths, and characteristics of the components of complex, a kind of intelligent concrete defect nondestructive detection algorithm based on multi resolution singular entropy was put forward. By using the wavelet algorithm, the ultrasonic signal was decomposed into high, low frequency components of multi scales for each component, then each component was decomposed by singular spectrum analysis, at the same time using the information entropy theory, the calculation of singular entropy as the characteristic defect detection value, using the GA-SVM algorithm was used to train the singular entropy, so as to achieve the purpose of identification of defects in concrete. The simulation results show that the proposed method can effectively improve the accuracy of the identification of the defects, and it can provide a new effective way to identify the defects of the concrete.

关 键 词：小波变换 奇异谱分析 信息熵 遗传算法 支持向量机 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]