太赫兹光谱在材料缺陷检测以及分类识别方面的应用

   太赫兹波是波段位于微波与红外线之间电磁波的统称。因其独特的性质，引起了研究人员的广泛关注。其中，太赫兹时域光谱技术（THz-TDS）是太赫兹波的一项重要应用，与传统光谱检测技术相比，具有高效率、安全性、稳定性等优势。但是系统中以热噪声和散粒噪声为主的不确定因素的存在，阻碍了THz-TDS系统进一步商业化精细仪器化发展，限制了THz-TDS系统的应用范围。因此，亟需针对太赫兹时域信号的降噪方法，得到高信噪比的太赫兹时域信号，从而使THz-TDS系统获得更广泛应用。

   经验模态分解（EMD）在信号降噪领域适用性很强，在非平稳信号降噪实践中效果显著，但是其中存在的模态混叠问题限制了该降噪算法的效果。本研究针对EMD中存在的模态混叠问题做了改进，将EMD与软阈值过滤（ST）、独立成分分析（ICA）相结合，提出新的降噪算法（EMD-ST-ICA），经过仿真验证，该算法降噪效果明显优于目前常用的几种降噪算法。

  经验小波变换（EWT）是在EMD的思想基础上发展出的一种信号分解方法。在太赫兹信号领域，尚无研究使用该信号分解方法进行降噪研究。本研究将EWT与小波阈值降噪（WT）相结合，提出经验小波变换-小波降噪（EWT-WT）的降噪算法。经过仿真信号与实际信号验证，该方法降噪效果显著优于目前常用的几种降噪算法。

  在得到了高信噪比的太赫兹信号后，本文继续对THz-TDS的实际应用进行了研究，分别使用THz-TDS系统对纤维增强聚合物复合材料内部的缺陷进行了检测和对沙土含水量自动分类识别。结果显示，在材料内部缺陷检测和沙土含水量分类识别实验中，THz-TDS系统均具有不可替代的应用潜力。

[1] ABBOTT D, ZHANG X-C. Special Issue on T-Ray Imaging, Sensing, and Retection [J]. Proceedings of the IEEE, 2007, 95(8): 1509-13.
[2] HAN P Y, TANI M, USAMI M, et al. A direct comparison between terahertz time-domain spectroscopy and far-infrared Fourier transform spectroscopy [J]. Journal of Applied Physics, 2001, 89(4): 2357-9.
[3] JULIANO T R, JR., KORTER T M. Terahertz vibrations of crystalline acyclic and cyclic diglycine: benchmarks for London force correction models [J]. J Phys Chem A, 2013, 117(40): 10504-12.
[4] NIIJIMA S, SHOYAMA M, MURAKAMI K, et al. Evaluation of the sintering properties of pottery bodies using terahertz time-domain spectroscopy [J]. Journal of Asian Ceramic Societies, 2018, 6(1): 37-42.
[5] SMOLYANSKAYA O A, CHERNOMYRDIN N V, KONOVKO A A, et al. Terahertz biophotonics as a tool for studies of dielectric and spectral properties of biological tissues and liquids [J]. Progress in Quantum Electronics, 2018, 62: 1-77.
[6] PELLIZZERI S, DELANEY S P, KORTER T M, et al. Using solid-state density functional theory and terahertz spectroscopy to spectroscopically distinguish the various hydrohalide salts of 5-(4-pyridyl)tetrazole [J]. Journal of Molecular Structure, 2013, 1050: 27-34.
[7] MAAMAR N, LAZOUL M, LATRECHE F Y, et al. Terahertz time-domain spectroscopy characterization of nitrocellulose in transmission and reflection configurations [J]. Optik, 2020, 224.
[8] IBRAHIM M E, HEADLAND D, WITHAYACHUMNANKUL W, et al. Nondestructive Testing of Defects in Polymer–Matrix Composite Materials for Marine Applications Using Terahertz Waves [J]. Journal of Nondestructive Evaluation, 2021, 40(2).
[9] PENKOV N V, GOLTYAEV M V, ASTASHEV M E, et al. The Application of Terahertz Time-Domain Spectroscopy to Identification of Potato Late Blight and Fusariosis [J]. Pathogens, 2021, 10(10).
[10] FITZGERALD A J, PINDER S, PURUSHOTHAM A D, et al. Classification of terahertz-pulsed imaging data from excised breast tissue [J]. J Biomed Opt, 2012, 17(1): 016005.
[11] YANG X, WU T, ZHANG L, et al. CNN with spatio-temporal information for fast suspicious object detection and recognition in THz security images [J]. Signal Processing, 2019, 160: 202-14.
[12] SERPENGüZEL A, WITHAYACHUMNANKUL W, BADENES G, et al. Analysis of measurement uncertainty in THz-TDS [Z]. Photonic Materials, Devices, and Applications II. 2007.10.1117/12.721876
[13] WU W, PENG H. Application of EMD Denoising Approach in Noisy Blind Source Separation [J]. Journal of Communications, 2014, 9(6): 506-14.
[14] HUANG N E, SHEN Z, LONG S R, et al. The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis [J]. Proceedings of the Royal Society of London Series A: Mathematical, Physical and Engineering Sciences, 1998, 454(1971): 903-95.
[15] JIANG X, WU F, YU H, et al. Mixed pixel decomposition of mineral spectrum based on EMD-ICA method [J]. Optics and Spectroscopy, 2015, 119(5): 893-8.
[16] GILLES J. Empirical Wavelet Transform [J]. IEEE Transactions on Signal Processing, 2013, 61(16): 3999-4010.
[17] BODA S, MAHADEVAPPA M, DUTTA P K. A hybrid method for removal of power line interference and baseline wander in ECG signals using EMD and EWT [J]. Biomedical Signal Processing and Control, 2021, 67.
[18] SADIQ M T, YU X, YUAN Z, et al. Motor Imagery EEG Signals Classification Based on Mode Amplitude and Frequency Components Using Empirical Wavelet Transform [J]. IEEE Access, 2019, 7: 127678-92.
[19] KEDADOUCHE M, LIU Z, VU V-H. A new approach based on OMA-empirical wavelet transforms for bearing fault diagnosis [J]. Measurement, 2016, 90: 292-308.
[20] HE M, NIAN Y, XU L, et al. Adaptive Separation of Respiratory and Heartbeat Signals among Multiple People Based on Empirical Wavelet Transform Using UWB Radar [J]. Sensors (Basel), 2020, 20(17).
[21] AJITO K, UENO Y. THz Chemical Imaging for Biological Applications [J]. IEEE Transactions on Terahertz Science and Technology, 2011, 1(1): 293-300.
[22] YANG X, ZHAO X, YANG K, et al. Biomedical Applications of Terahertz Spectroscopy and Imaging [J]. Trends Biotechnol, 2016, 34(10): 810-24.
[23] ZHANG L, ZHONG H, DENG C, et al. Terahertz wave reference-free phase imaging for identification of explosives [J]. Applied Physics Letters, 2008, 92(9).
[24] LEE G-J, KIM S, KWON T-H. Effect of Moisture Content and Particle Size on Extinction Coefficients of Soils Using Terahertz Time-Domain Spectroscopy [J]. IEEE Transactions on Terahertz Science and Technology, 2017, 7(5): 529-35.
[25] WANG Q, ZHOU H, LIU M, et al. Study of the skin depth and defect detection in carbon fiber composites with Terahertz waves [J]. Optik, 2019, 178: 1035-44.
[26] ZHIDONG C, LIJUN Y, LI C, et al. Application of Terahertz Spectroscopy in Defect Detection of Power Cable Terminal Insulation Material [Z]. 2021 46th International Conference on Infrared, Millimeter and Terahertz Waves (IRMMW-THz). 2021: 1-2.10.1109/IRMMW-THz50926.2021.9567419
[27] SERPENGüZEL A, WITHAYACHUMNANKUL W, BADENES G, et al. Analysis of measurement uncertainty in THz-TDS [Z]. Photonic Materials, Devices, and Applications II. 2007.10.1117/12.721876.
[28]李琦,夏志伟,丁胜晖,王骐.采用非局部均值的连续太赫兹图像去噪处理[J].红外与激光工程,2012,41(02):517-522.
[29] DU X, FAN D, LE J, et al. Terahertz digital holography image denoising using stationary wavelet transform [Z]. Selected Papers from Conferences of the Photoelectronic Technology Committee of the Chinese Society of Astronautics 2014, Part II. 2015.10.1117/12.2182877
[30] WANG S, NIU P, GUO Q, et al. An adaptive empirical mode decomposition and stochastic resonance system in high efficient detection of terahertz radar signal [J]. Ferroelectrics, 2020, 563(1): 148-60.
[31] NUNES J C, GUYOT S, DEL CHELLE E. Texture analysis based on local analysis of the Bidimensional Empirical Mode Decomposition [J]. Machine Vision and Applications, 2005, 16(3): 177-88.
[32] WU W, PENG H. Application of EMD Denoising Approach in Noisy Blind Source Separation [J]. Journal of Communications, 2014, 9(6): 506-14.
[33] QIAO X, ZHANG X, REN J, et al. Mean estimation empirical mode decomposition method for terahertz time-domain spectroscopy de-noising [J]. Appl Opt, 2017, 56(25): 7138-45.
[34] QIN Z, CHEN H, CHANG J. Signal-to-Noise Ratio Enhancement Based on Empirical Mode Decomposition in Phase-Sensitive Optical Time Domain Reflectometry Systems [J]. Sensors (Basel), 2017, 17(8).
[35] XIAO F, YANG D, GUO X, et al. VMD-based denoising methods for surface electromyography signals [J]. J Neural Eng, 2019, 16(5): 056017.
[36] ZHOU M, ZHAO H, XIA H, et al. De-noising of photoacoustic sensing and imaging based on combined empirical mode decomposition and independent component analysis [J]. J Biophotonics, 2019, 12(8): e201900042.
[37] CHANCHANG V, KUMCHAISEEMAK N, SUTTHIOPAD M, et al. Improvement of electrocardiogram by empirical wavelet transform [J]. Journal of Physics: Conference Series, 2017, 901.
[38] HU Y, LI F, LI H, et al. An enhanced empirical wavelet transform for noisy and non-stationary signal processing [J]. Digital Signal Processing, 2017, 60: 220-9.
[39] CHEN W, SONG H. Automatic noise attenuation based on clustering and empirical wavelet transform [J]. Journal of Applied Geophysics, 2018, 159: 649-65.
[40]陈学军,杨永明.采用经验小波变换的风力发电机振动信号消噪[J].浙江大学学报(工学版),2018,52(05):988-995.
[41] SINGH O, SUNKARIA R K. ECG signal denoising via empirical wavelet transform [J]. Australas Phys Eng Sci Med, 2017, 40(1): 219-29.
[42] MAHESHWARI S, PACHORI R B, ACHARYA U R. Automated Diagnosis of Glaucoma Using Empirical Wavelet Transform and Correntropy Features Extracted From Fundus Images [J]. IEEE J Biomed Health Inform, 2017, 21(3): 803-13.
[43] SIULY, YIN X, HADJILOUCAS S, et al. Classification of THz pulse signals using two-dimensional cross-correlation feature extraction and non-linear classifiers [J]. Comput Methods Programs Biomed, 2016, 127: 64-82.
[44] LIU W, ZHANG R, LING Y, et al. Automatic recognition of breast invasive ductal carcinoma based on terahertz spectroscopy with wavelet packet transform and machine learning [J]. Biomed Opt Express, 2020, 11(2): 971-81.
[45] CACCIARI I, CORRADI G. Common plastics THz classification via artificial neural networks: A discussion on a class of time domain features [J]. Optical Materials, 2021, 117.
[46] NEU J, SCHMUTTENMAER C A. Tutorial: An introduction to terahertz time domain spectroscopy (THz-TDS) [J]. Journal of Applied Physics, 2018, 124(23).
[47] ZHANG J, WANG J, HAN X, et al. Noncontact detection of Teflon inclusions in glass-fiber-reinforced polymer composites using terahertz imaging [J]. Appl Opt, 2016, 55(36): 10215-22.
[48] ZHANG Y, ZHANG X, LI S, et al. A Broadband THz-TDS System Based on DSTMS Emitter and LTG InGaAs/InAlAs Photoconductive Antenna Detector [J]. Sci Rep, 2016, 6: 26949.
[49] NAFTALY M. Metrology Issues and Solutions in THz Time-Domain Spectroscopy: Noise, Errors, Calibration [J]. IEEE Sensors Journal, 2013, 13(1): 8-17.
[50]潘武,刘博文,马勇,肖惠云,杨龙亮.偏振可调的太赫兹光电导天线设计[J].半导体光电,2022,43(01):110-115.
[51]Pupeza Ioachim,Wilk Rafal,Koch Martin. Highly accurate optical material parameter determination with THz time-domain spectroscopy.[J]. Optics express,2007,15(7).
[52] 李雅卓,李向军,洪治.基于太赫兹时域光谱的半导体材料参数测量[J].半导体技术,2008,33(12):1074-1076+1099.
[53] WANG R, SUN S, GUO X, et al. EMD Threshold Denoising Algorithm Based on Variance Estimation [J]. Circuits, Systems, and Signal Processing, 2018, 37(12): 5369-88.
[54] INAGAKI T, AHMED B, HARTLEY I D, et al. Simultaneous prediction of density and moisture content of wood by terahertz time domain spectroscopy [J]. Journal of Infrared, Millimeter, and Terahertz Waves, 2014, 35(11): 949-61.
[55] LJUBENOVIC M, ZHUANG L, DE BEENHOUWER J, et al. Joint Deblurring and Denoising of THz Time-Domain Images [J]. IEEE Access, 2021, 9: 162-76.
[56] OH G-H, KIM H-S, PARK D-W, et al. In-situ monitoring of moisture diffusion process for wood with terahertz time-domain spectroscopy [J]. Optics and Lasers in Engineering, 2020, 128.
[57] GARG D, BANDYOPADHYAY A, SENGUPTA A. Measurement of moisture content in milk powder using terahertz time-domain spectroscopy [M]. SPIE, 2021