薄厚度有机防护涂层的太赫兹无损检测与多元回归分析

Regression Analysis in Terahertz Quantitative Nondestructive Evaluation of Thinner Organic Protective Coatings

传统太赫兹检测数据处理方法中,介质层最小检测厚度受限于太赫兹在介质中的脉冲相干长度的一半。采用引入粒子群优化的多元回归分析方法来提高太赫兹检测有机防护涂层厚度分辨率。分别以太赫兹原始检测信号、傅里叶反卷积得到的脉冲响应函数、信号经平稳小波变换分解的细节系数为对象,提取不同特征参数建立相应的多元回归模型,并引入粒子群算法进行优化,最后用相应模型预测薄厚度涂层。预测分析结果表明:采用本文方法对太赫兹检测信号的平稳小波变换细节系数进行分析,能够有效提高太赫兹脉冲对薄厚度涂层的检测精度,有利于涂层涂装生产中的质量监控。

In the traditional terahertz(THz)detection data processing methods,the minimum thickness is constrained to be half of the coherent length of the THz pulse in a sample.A novel approach is presented to enhance the capability of resolving thinner organic coating layers using terahertz pulsed imaging(TPI)’s based on multivariate regression analysis combined with particle swarm optimization(PSO)approach.The THz original detected signals,the impulse functions obtained by Fourier deconvolution,and the detail coefficient decomposed by Stationary Wavelet Transform(SWT)method are preprocessed for extracting regressive parameters.The different multiple regressive models are established combined with PSO algorithm,which are adopted to predict the thinner coating layer.The results showed that the proposed method with SWT detail coefficients of TPI signal can effectively improve the detection precision of the thin thickness coating by THz pulse,and is advantageous to the quality control method in the coating production.