Quantitative analysis of terahertz signals using CWT-based spectrogram and Zernike image moments.

The combination of terahertz (THz) spectroscopic measurements and multivariate calibration techniques has become a well-established technique in many research fields. However, intentional or unintentional changes in environmental conditions, THz instruments and/or of the substance itself make the established calibration model becoming insufficient and inadequate for the further application. In this article, we introduce, discuss, and evaluate a new multivariate calibration method, the CWT-ZM, that combines the merits of the Zernike moment (ZM) invariance and the continuous wavelet transform (CWT) time-frequency analysis. With the help of a wavelet time-frequency analysis, the THz pulse is expanded into a two-dimensional (2D) time-frequency plane that provides richer and more direct characteristic information in the time and frequency domain simultaneously. In addition, Zernike moments provide linearly independent descriptors for the 2D time-frequency intensity image and are invariant to THz signal affine transformations, such as peak shifting, baseline drifting, and scaling. In this manner, we obtain a set of features that exhibit a high capability to capture the concentrations of the target compounds and a high invariance of the different measuring instruments and the variable environment. This approach results in a more robust regression system with improved generalization properties with respect to standard methods. Experiments were then conducted on a THz dataset of pharmaceutical tablets acquired by two different THz instruments, and these confirmed the effectiveness of the proposed approach. Furthermore, CWT-ZM is an extensible framework that can be combined with various spectral qualitative and quantitative analysis algorithms.

Rapid Identification of Soybean Varieties by Terahertz Frequency-Domain Spectroscopy and Grey Wolf Optimizer-Support Vector Machine.

Different soybean varieties vary greatly in their nutritional value and composition. Screening for superior varieties is also essential for the development of the soybean seed industry. The objective of the paper was to analyze the feasibility of terahertz (THz) frequency-domain spectroscopy and chemometrics for soybean variety identification. Meanwhile, a grey wolf optimizer-support vector machine (GWO-SVM) soybean variety identification model was proposed. Firstly, the THz frequency-domain spectra of experimental samples (6 varieties, 270 in total) were collected. Principal component analysis (PCA) was used to analyze the THz spectra. After that, 203 samples from the calibration set were used to establish a soybean variety identification model. Finally, 67 samples from the test set were used for prediction validation. The experimental results demonstrated that THz frequency-domain spectroscopy combined with GWO-SVM could quickly and accurately identify soybean varieties. Compared with discriminant partial least squares (DPLS) and particles swarm optimization support vector machine, GWO-SVM combined with the second derivative could establish a better soybean variety identification model. The overall correct identification rate of its prediction set was 97.01%.

Terahertz signal analysis and substance identification via Zernike moments.

Terahertz (THz) spectra contain chemical information, along with noise and variable backgrounds. Measurement environmental changes and spectral signal differences caused by changes in the sample state can degrade the accuracy of the calibration models. This problem obviously hinders practical applications of THz spectroscopy. To tackle this problem, a three-dimensional spectrum was first self-constructed and converted into an intensity image. Zernike moments with inherently invariant properties were then used to describe the THz intensity image and extract the invariant features for further analysis. Considering the reconstruction error and computational cost, the highest order of Zernike moments and the most effective moments were selected and applied to multi-classifiers including support vector machines, naive Bayes, and regularized linear discriminant analysis. Experiments used a THz dataset collected from four chemical substances (melamine, tartaric acid, lactose, and glucose) at five thicknesses (1.0 mm, 1.5 mm, 2.0 mm, 2.5 mm, and 3.0 mm). The results confirmed the effectiveness of the proposed approach. The obtained results show that compare to traditional absorption spectrum features, Zernike moment features are less sensitive to spectral variations caused by changes in sample status. They have better feature representation ability with lower feature vector dimensions. This suggests that they can be integrated into the design of systems for THz spectral classification to increase the robustness and generalization capability of the classifier.

Development of a mechanism for reconstruction of terahertz single-frequency images of biological samples.

Algorithmic mechanisms are used to improve terahertz (THz) image quality, which is critical to a biological sample analysis. A complete mechanism for the super-resolution reconstruction and evaluation of THz biological sample images was constructed in this study. With eucalyptus leaves as an example, the THz spectral region screening technique was adopted to select the characteristic frequencies for imaging, and the THz single-frequency images were reconstructed with the single-image super-resolution image reconstruction technique. The THz super-resolution reconstructed images without ideal reference were evaluated after the introduction of three no-reference image evaluation criteria considering the diversity and complexity of organisms. The results show that the THz image reconstruction mechanism proposed in this study led to an increase in resolution and a decrease in noise. At the same time, the imaging quality of biological samples was considerably improved, and the detailed information was enriched. These provide a reference for a THz imaging analysis of leaves and other biological samples.

Terahertz, infrared and Raman absorption spectra of tyrosine enantiomers and racemic compound.

The application of terahertz (THz)-based techniques in biomolecule study is very promising but still in its infancy. In the present work, we employed THz time-domain spectroscopy (THz-TDS) and THz time-domain attenuated total reflection (THz-TD-ATR) spectroscopy to investigate the properties of tyrosine (Tyr) enantiomers (L- and D-Tyr) and racemate (DL-Tyr) in solid state and aqueous solutions, respectively. THz absorption spectra of solid L- and D-Tyr show similar absorption spectra with peaks at 0.95, 1.92, 2.06 and 2.60 THz, which are obviously different from the spectrum of DL-Tyr with peaks at 1.5, 2.15 and 2.40 THz. In contrast, although THz absorption spectra of L-Tyr solution and D-Tyr solution are similar and different from the spectrum of DL-Tyr solution, both of them have no observable peaks. Interestingly, it was found that the solution containing equal amounts of L- and D-Tyr has a similar spectrum as that of DL-Tyr solution, as far as the mass concentrations of the two types of solutions are kept the same. On other hand, solid L-, D- and DL-Tyr were also investigated with infrared spectroscopy and Raman spectroscopy, respectively. The results show that the spectra of L- and D-Tyr can be regarded the same but they are slightly different from the spectrum of DL-Tyr. With the aid of principal component analysis (PCA), the difference between L-/D-Tyr and DL-Tyr can be confirmed without any ambiguity. Overall, this work systematically interrogated and evaluated the performance of THz-based techniques in the detection of the chirality of tyrosine.

Terahertz spectroscopy combined with data dimensionality reduction algorithms for quantitative analysis of protein content in soybeans.

Protein content in soybean is a key determinant of its nutritional and economic value. The paper investigated the feasibility of terahertz (THz) spectroscopy and dimensionality reduction algorithms for the determination of protein content in soybean. First of all, the THz sample spectrum was data processed by pre-processing or dimensionality reduction algorithms. Secondly, by calibration set, using partial least squares regression (PLSR), genetic algorithms-support vector regression (GA-SVR), grey wolf optimizer-support vector regression (GWO-SVR) and back propagation neural network (BPNN) were respectively used to model protein content determination. Afterwards, the model was validated by the prediction set. Ultimately, the BPNN model combined with linear discriminant analysis (LDA) for related coefficient of prediction set (Rp), root mean square error of prediction set (RMSEP), relative standard deviation (RSD), the time required for the operation was respectively 0.9677, 1.2467%, 3.3664%, and 53.51 s. The experimental results showed that the rapid and accurate quantitative determination of protein in soybean using THz spectroscopy is feasible after a suitable dimensionality reduction algorithm.

Application of terahertz spectrum and interval partial least squares method in the identification of genetically modified soybeans.

Genetically modified soybeans are the world's most important genetically modified agricultural product. At present, the traditional methods for identifying genetically modified and non-transgenic soybeans are time-consuming, costly, and complicated to operate, which cannot meet the needs of practical applications. Therefore, it is necessary to discover a fast and accurate method for identifying transgenic soybeans. Terahertz (THz) time domain spectra were collected in sequence from 225 transgenic and non-transgenic soybean samples. Fourier transform was used to convert the terahertz time domain spectrum into a THz frequency domain spectrum with a frequency range of 0.1-2.5 THz. Firstly, the interval partial least squares (iPLS) method was used to remove interference spectral bands and select appropriate spectral intervals. Secondly, 168 samples were selected as the calibration set. Discriminant partial least squares (DPLS), Grid Search support vector machine (Grid Search-SVM) and principal component analysis back propagation neural network (PCA-BPNN) were used to establish a qualitative identification model. Afterwards, 57 test set samples were predicted. By comparing the experimental results, it was found that iPLS could effectively screen and remove the interference THz band, which was more helpful to improve the efficiency and accuracy of the identification model. After the iPLS and mean center pre-treatment technology, the Grid Search-SVM identification model had the best identification effect, with a total accuracy rate of 98.25% (transgenic identification rate was 96.15%, non-transgenic identification rate was 100%). This study shows that after selecting spectra from iPLS, THz spectroscopy combined with chemometrics can more accurately, quickly, and efficiently identify transgenic and non-transgenic soybeans.

Imaging brain tissue slices with terahertz near-field microscopy.

Photoconductive antenna microprobe (PCAM)-based terahertz (THz) near-field imaging technique is promising for biomedical detection due to its excellent biocompatibility and high resolution; yet it is limited by its imaging speed and the difficulty in the control of the PCAM tip-sample separation. In this work, we successfully realized imaging of mouse brain tissue slices using an improved home-built PCAM-based THz near-field microscope. In this system, the imaging speed was enhanced by designing and applying a voice coil motor-based delay-line. The tip-sample separation control was implemented by developing an image analysis-based technique. Compared with conventional PCAM-based THz near-field systems, our improved system is 100 times faster in imaging speed and the tip-sample separation can be controlled to a few micrometers (e.g., 3 µm), satisfying the requirements of THz near-field imaging of biological samples. It took about ~30 min (not the tens of hours it took to acquire the same kind of image previously) to collect a THz near-field image of brain tissue slices of BALb/c mice (500 µm × 500 µm) with pixel size of 20 µm × 20 µm. The results show that the mouse brain slices can be properly imaged and different regions in the slices (i.e., the corpus callosum region and the cerebrum region) can be identified unambiguously. Evidently, the work demonstrated here provides not only a convincing example but a useful technique for imaging biological samples with THz near-field microscopy. © 2018 American Institute of Chemical Engineers Biotechnol. Prog., 35: e2741, 2019.

Terahertz Time Domain Spectroscopy of Transformer Insulation Paper after Thermal Aging Intervals.

An accelerated thermal aging process was used to simulate the condition of paper insulation in transformer oil-paper systems. Optical parameters of the insulation paper after various aging intervals were analyzed with terahertz time-domain spectroscopy (THz-TDS) over the range 0.1~1.8 THz. The result shows that the paper had seven absorption peaks at 0.19, 0.49, 0.82, 1.19, 1.43, 1.53, and 1.74 THz, and density functional theory of B3LYP/6-311G+ (d, p) was used to simulate the molecular dynamics of the repeating component (cellobiose) of the cellulose paper. Theoretical spectra were consistent with experiment, which had absorption peaks at 0.18, 0.82, 1.47, and 1.53 THz in the same frequency range. At the same time, the paper samples after various aging intervals had different refractive indexes, and least squares fitting revealed a linear relationship between the degree of polymerization and the refractive index of the paper. Hence, this paper demonstrates that THz-TDS could be used to analyze the aging condition of transformer insulation paper and provides the theoretical and experimental basis for detection.

Rapid discrimination of Notoginseng powder adulteration of different grades using FT-MIR spectroscopy combined with chemometrics.

Panax Notoginseng is a kind of herb material with high medicinal value, which requires adaptive planting environment, and not can be continuously cultivated in the same ground. Those reasons lead to a large number of low-grade Notoginseng appears in the market. The objective of this study is to discriminate adulterant of Notoginseng of different grades by FT-MIR spectroscopy couple with chemometrics. In the experiment, high-grade Notoginseng was adulterated with 14 blend ratios: 0%, 1%, 3%, 5%, 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, and 100% of low-grade Notoginseng. All samples were scanned in the range of 4000-400 cm-1 by FT-MIR spectra instrument in absorption mode. Baseline, standard normal variate (SNV), multiplicative scatter correction (MSC), orthogonal signal correction (OSC), first derivative (D1) with 11-points smoothing and second derivative (D2) with 11-points smoothing were used to preprocess the spectral data, in which Baseline combined with SNV and D1 with 11-points performed best. The spectral data in the range of 1485-405 cm-1 were selected by interval partial least squares (iPLS) for modeling. Then, Support vector machine (SVM) and linear discriminant analysis (LDA) were applied for modeling analysis. The best result was achieved by SVM, as the classification accuracy was 100%, which indicated that FT-MIR spectroscopy combined with chemometrics was an effective approach to identify Notoginseng powder adulteration. It could detect the blend ratio of 5% (w/w) as well as the blend ratio of over 5%.