基于修正自相关算法的THz-TDS聚乙烯测厚反卷积算法  

Deconvolution Algorithm of THz-TDS Polyethylene Thickness Measurement Based on Modified Autocorrelation Algorithm

作　　者：蒋强 俞跃[3] 叶凌伟 郝元[4] JIANG Qiang;YU Yue;YE Lingwei;HAO Yuan(Zhejiang Special Equipment Science Research Institute,Hangzhou 341000,China;Zhejiang Key Laboratory of Special Equipment Safety Testing Technology Research,Hangzhou 341000,China;China Special Equipment Inspection Institute,Beijing 100029,China;China Measurement University,Zhejiang 310018,China)

机构地区：[1]浙江省特种设备科学研究院,浙江杭州310018 [2]浙江省特种设备安全检测技术研究重点实验室,浙江杭州3100183 [3]中国特种设备检测研究院,北京100029 [4]中国计量大学,浙江杭州310018

出　　处：《红外技术》2020年第5期473-482,共10页Infrared Technology

基　　金：国家重点研发计划项目(2017YFF0209704)。

摘　　要：在聚乙烯材料管道生产制造过程中,管道厚度精确测量是影响管道生产质量的一个重要技术问题,太赫兹非金属厚度透射测量过程中,通过测量两次透射信号的光程时间差与材料折射率计算得到样品厚度参数,为了准确提取两次透射信号的光程时间差值,需要对原始时域信号进行信号表征。通过高斯滤波反卷积得到脉冲响应函数表征时域信号已经成为了一种成熟且有效的技术手段。但一方面由于太赫兹波在聚乙烯材料中的传播存在明显的频散现象,同时其透射率高回波信号较弱,导致信噪比较低,在脉冲响应函数提取过程中放大了干扰信号的比重,造成信号失真、淹没,无法准确提取信号。本文通过对带通滤波后的时域信号采用修正型自相关算法进行自相关性分析,再结合高斯滤波反卷积运算,得到改进后的脉冲响应函数。旨在解决太赫兹非金属测量领域时域信号表征问题。改进算法实验结果显示加强了信号的清晰度,解决了因杂波信号干扰造成的脉冲响应信号失真或淹没情况,显著提高了脉冲响应信号信噪比。In polyethylene pipe manufacturing processes,the precise measurement of the pipe thickness is an important technical problem,which affects the quality of the pipes produced.In the terahertz non-metal thickness transmission measurement process,the sample thickness parameter is obtained by measuring the optical path time difference of the two transmitted signals and material refractive index.To accurately extract the optical path time difference of the two transmitted signals,signal characterization of the original time domain signal is required,which is a mature and effective technical means to characterize the signals impulse response functions.An impulse response function can be obtained using a deconvolution technique.However,when the echo signal is weak or the system signal-to-noise ratio is low,the interference signal increases,causing distortion and flooding of the signal,and thus,the signal cannot be accurately extracted.In this study,we first pre-process the time domain signal with a Butterworth filter and perform an improved self-precision analysis on the processed signal.Second,we perform Gaussian deconvolution processing for autocorrelation signals.Finally,we obtain an improved impulse response function.The improved method aims to solve the problem of signal characterization in the field of terahertz non-metal measurement.The improved algorithm enhances the clarity of the signal and solves the distortion or flooding of the impulse response signal caused by the interference of the clutter signal.Consequently,the signal-to-noise ratio of the impulse response signal is improved.

关 键 词：太赫兹 反卷积 信号表征 自相关函数 信噪比 

分 类 号：TB324[一般工业技术—材料科学与工程]