摘要
针对传统金属材料缺陷检测存在的弊端,结合当前的电磁层析成像技术,提出一种基于L1的正则化图像重建方案。在该方案中,根据电磁层析成像技术逆问题原理,在构建电磁层析成像数学描述的基础上,采用稀疏表示算法对模型进行求解,从而得到测量数据向量、灵敏度矩阵、图像灰度值向量三者之间的关系。为验证上述算法的可行性,从传感器设计、激励信号方案、采集模块等方面对电磁层析成像系统进行构建,并以铝盘作为被测对象。结果表明,采用L1的正则化得到的图像与真实金属材料缺陷分布更加接近。
Aiming at the drawbacks of traditional metal material defect detection,a regularized image reconstruction scheme based on L1 is pro.posed in combination with current EMT technology.In this scheme,based on the principle of inverse problem of EMT technology and the con.struction of mathematical description of EMT,the sparse representation algorithm is used to solve the model,and the relationship among mea.surement data vector,sensitivity matrix and image gray value vector are obtained.In order to verify the feasibility of the above algorithm,the EMT system is constructed from the aspects of sensor design,excitation signal scheme,acquisition module,and the aluminium disc is taken as the tested object.The results show that the image obtained by L1 regularization is closer to the defect distribution of real metal materials.
作者
许大伟
XU Dawei(Xi’an Aeronautical Polytechnic Institute,Xi’an 710089,China)
出处
《工业加热》
CAS
2019年第3期74-77,共4页
Industrial Heating
关键词
L1正则化
金属材料
缺陷检测
L1 regularization
metal materials
defect detection
