Multi-variable selection strategy based on near-infrared spectra for the rapid description of dianhong black tea quality

Detecting bioactive compound contents in Dancong tea using VNIR-SWIR hyperspectral imaging and KRR model with a refined feature wavelength method

Hyperspectral imaging (HSI) provides opportunity for non-destructively detecting bioactive compounds contents of tea leaves and high detection accuracy require extracting effective features from the complex hyperspectral data. In this paper, we proposed a feature wavelength refinement method called interval band selecting-competitive adaptive reweighted sampling-fusing (IBS-CARS-Fusing) to extract feature wavelengths from visible-near-infrared (VNIR) and short-wave-near-infrared (SWIR) hyperspectral images. Combined with the proposed IBS-CARS-Fusing method, a kernel ridge regression (KRR) model was established to predict the contents of bioactive compounds including chlorophyll a, chlorophyll b, carotenoids, tea polyphenols, and amino acids in Dancong tea. It was revealed that the IBS-CARS-Fusing method can improve Rp2 of KRR model for these bioactive compounds by 4.77%, 4.60%, 6.74%, 15.52%, and 13.10%, respectively, and Rp2 of the model reached high values of 0.9500, 0.9481, 0.8946, 0.8882, and 0.8622. Additionally, a leaf compound mass per area thermal map was used to visualize the spatial distribution of the compounds.

Simultaneous Determination of Four Catechins in Black Tea via NIR Spectroscopy and Feature Wavelength Selection: A Novel Approach

As a non-destructive, fast, and cost-effective technique, near-infrared (NIR) spectroscopy has been widely used to determine the content of bioactive components in tea. However, due to the similar chemical structures of various catechins in black tea, the NIR spectra of black tea severely overlap in certain bands, causing nonlinear relationships and reducing analytical accuracy. In addition, the number of NIR spectral wavelengths is much larger than that of the modeled samples, and the small-sample learning problem is rather typical. These issues make the use of NIRS to simultaneously determine black tea catechins challenging. To address the above problems, this study innovatively proposed a wavelength selection algorithm based on feature interval combination sensitivity segmentation (FIC-SS). This algorithm extracts wavelengths at both coarse-grained and fine-grained levels, achieving higher accuracy and stability in feature wavelength extraction. On this basis, the study built four simultaneous prediction models for catechins based on extreme learning machines (ELMs), utilizing their powerful nonlinear learning ability and simple model structure to achieve simultaneous and accurate prediction of catechins. The experimental results showed that for the full spectrum, the ELM model has better prediction performance than the partial least squares model for epicatechin (EC), epicatechin gallate (ECG), epigallocatechin (EGC), and epigallocatechin gallate (EGCG). For the feature wavelengths, our proposed FIC-SS-ELM model enjoys higher prediction performance than ELM models based on other wavelength selection algorithms; it can simultaneously and accurately predict the content of EC (Rp2 = 0.91, RMSEP = 0.019), ECG (Rp2 = 0.96, RMSEP = 0.11), EGC (Rp2 = 0.97, RMSEP = 0.15), and EGCG (Rp2 = 0.97, RMSEP = 0.35) in black tea. The results of this study provide a new method for the quantitative determination of the bioactive components of black tea.

Quantitative analysis of aflatoxin B1 of peanut by optimized support vector machine models based on near-infrared spectral features

This study designs a chemometric framework for quantitatively evaluating aflatoxin B1 (AFB1) in peanuts based on near-infrared (NIR) spectroscopy technique. The NIR spectra of peanut samples exhibiting diverse fungal contamination levels were acquired using a portable NIR spectrometer. Subsequently, appropriate pre-processing techniques were employed for data refinement. To streamline the analysis, the iterative variable subset optimization (IVSO) technique was employed to conduct an initial screening of the pre-processed NIR spectra, eliminating numerous irrelevant variables. Building upon this screening process, the beluga whale optimization (BWO) algorithm was utilized to optimize the selected feature variables further. Subsequently, support vector machine (SVM) models were developed using the refined near-infrared spectral features to test AFB1 in peanuts quantitatively. The results indicate that the SVM model significantly improves detection performance and generalization proficiency, particularly after secondary optimization using BWO-IVSO. Among the different models considered, the SVM model established after BWO-IVSO optimization exhibited the most extraordinary level of generalization, with a root mean square error of prediction of 24.6322 μg∙kg-1, a correlation coefficient of 0.9761, and a relative percent deviation of 4.6999. Overall, this investigation highlights the effectiveness of the proposed NIR spectroscopy model based on BWO-IVSO-SVM for quantitatively analyzing AFB1 in peanuts. The study contributes valuable technical and methodological insights that can serve as a reference for rapidly determining mycotoxins in cereal crops.

The Integration of Metabolomics, Electronic Tongue, and Chromatic Difference Reveals the Correlations between the Critical Compounds and Flavor Characteristics of Two Grades of High-Quality Dianhong Congou Black Tea

Tea's biochemical compounds and flavor quality vary depending on its grade ranking. Dianhong Congou black tea (DCT) is a unique tea category produced using the large-leaf tea varieties from Yunnan, China. To date, the flavor characteristics and critical components of two grades of high-quality DCT, single-bud-grade DCT (BDCT), and special-grade DCT (SDCT) manufactured mainly with single buds and buds with one leaf, respectively, are far from clear. Herein, comparisons of two grades were performed by the integration of human sensory evaluation, an electronic tongue, chromatic differences, the quantification of major components, and metabolomics. The BDCT possessed a brisk, umami taste and a brighter infusion color, while the SDCT presented a comprehensive taste and redder liquor color. Quantification analysis showed that the levels of total polyphenols, catechins, and theaflavins (TFs) were significantly higher in the BDCT. Fifty-six different key compounds were screened by metabolomics, including catechins, flavone/flavonol glycosides, amino acids, phenolic acids, etc. Correlation analysis revealed that the sensory features of the BDCT and SDCT were attributed to their higher contents of catechins, TFs, theogallin, digalloylglucose, and accumulations of thearubigins (TRs), flavone/flavonol glycosides, and soluble sugars, respectively. This report is the first to focus on the comprehensive evaluation of the biochemical compositions and sensory characteristics of two grades of high-quality DCT, advancing the understanding of DCT from a multi-dimensional perspective.

Identification of aged-rice adulteration based on near-infrared spectroscopy combined with partial least squares regression and characteristic wavelength variables

The long-term storage of rice will inevitably be involved in the deterioration of edible quality, and aged rice poses a great threat to food safety and human health. The acid value can be employed as a sensitive index for the determination of rice quality and freshness. In this study, near-infrared spectra of three kinds of rice (Chinese Daohuaxiang, southern japonica rice, and late japonica rice) mixed with different proportions of aged rice were collected. The partial least squares regression (PLSR) model with different preprocessing was constructed to identify the aged rice adulteration. Meanwhile, a competitive adaptive reweighted sampling (CARS) algorithm was used to extract the optimization model of characteristic variables. The constructed CARS-PLSR model method could not only reduce greatly the number of characteristic variables required by the spectrum but also improve the identification accuracy of three kinds of aged-rice adulteration. As above, this study proposed a rapid, simple, and accurate detection method for aged-rice adulteration, providing new clues and alternatives for the quality control of commercial rice.

NIR Spectrometric Approach for Geographical Origin Identification and Taste Related Compounds Content Prediction of Lushan Yunwu Tea

Lushan Yunwu Tea is one of a unique Chinese tea series, and total polyphenols (TP), free amino acids (FAA), and polyphenols-to-amino acids ratio models (TP/FAA) represent its most important taste-related indicators. In this work, a feasibility study was proposed to simultaneously predict the authenticity identification and taste-related indicators of Lushan Yunwu tea, using near-infrared spectroscopy combined with multivariate analysis. Different waveband selections and spectral pre-processing methods were compared during the discriminant analysis (DA) and partial least squares (PLS) model-building process. The DA model achieved optimal performance in distinguishing Lushan Yunwu tea from other non-Lushan Yunwu teas, with a correct classification rate of up to 100%. The synergy interval partial least squares (siPLS) and backward interval partial least squares (biPLS) algorithms showed considerable advantages in improving the prediction performance of TP, FAA, and TP/FAA. The siPLS algorithms achieved the best prediction results for TP (R_P = 0.9407, RPD = 3.00), FAA (R_P = 0.9110, RPD = 2.21) and TP/FAA (R_P = 0.9377, RPD = 2.90). These results indicated that NIR spectroscopy was a useful and low-cost tool by which to offer definitive quantitative and qualitative analysis for Lushan Yunwu tea.

Determination of Cultivation Regions and Quality Parameters of Poria cocos by Near-Infrared Spectroscopy and Chemometrics

Poria cocos (PC) is an important fungus with high medicinal and nutritional values. However, the quality of PC is heavily dependent on multiple factors in the cultivation regions. Traditional methods are not able to perform quality evaluation for this fungus in a short time, and a new method is needed for rapid quality assessment. Here, we used near-infrared (NIR) spectroscopy combined with chemometric method to identify the cultivation regions and determine PC chemical compositions. In our study, 138 batches of samples were collected and their cultivation regions were distinguished by combining NIR spectroscopy and random forest method (RFM) with an accuracy as high as 92.59%. In the meantime, we used partial least square regression (PLSR) to build quantitative models and measure the content of water-soluble extract (WSE), ethanol-soluble extract (ASE), polysaccharides (PSC) and the sum of five triterpenoids (SFT). The performance of these models were verified with correlation coefficients (R²_cal and R²_pre) above 0.9 for the four quality parameters and the relative errors (RE) of PSC, WSE, ASE and SFT at 4.055%, 3.821%, 4.344% and 3.744%, respectively. Overall, a new approach was developed and validated which is able to distinguish PC production regions, quantify its chemical contents, and effectively evaluate PC quality.

Spectral Diagnostic Model for Agricultural Robot System Based on Binary Wavelet Algorithm

The application of agricultural robots can liberate labor. The improvement of robot sensing systems is the premise of making it work. At present, more research is being conducted on weeding and harvesting systems of field robot, but less research is being conducted on crop disease and insect pest perception, nutritional element diagnosis and precision fertilizer spraying systems. In this study, the effects of the nitrogen application rate on the absorption and accumulation of nitrogen, phosphorus and potassium in sweet maize were determined. Firstly, linear, parabolic, exponential and logarithmic diagnostic models of nitrogen, phosphorus and potassium contents were constructed by spectral characteristic variables. Secondly, the partial least squares regression and neural network nonlinear diagnosis model of nitrogen, phosphorus and potassium contents were constructed by the high-frequency wavelet sensitivity coefficient of binary wavelet decomposition. The results show that the neural network nonlinear diagnosis model of nitrogen, phosphorus and potassium content based on the high-frequency wavelet sensitivity coefficient of binary wavelet decomposition is better. The R2, MRE and NRMSE of nn of nitrogen, phosphorus and potassium were 0.974, 1.65% and 0.0198; 0.969, 9.02% and 0.1041; and 0.821, 2.16% and 0.0301, respectively. The model can provide growth monitoring for sweet corn and a perception model for the nutrient element perception system of an agricultural robot, while making preliminary preparations for the realization of intelligent and accurate field fertilization.

Rapid identification of the green tea geographical origin and processing month based on near-infrared hyperspectral imaging combined with chemometrics

The geographical origin and processing month of green tea greatly affect its economic value and consumer acceptance. This study investigated the feasibility of combining near-infrared hyperspectral imaging (NIR-HSI) with chemometrics for the identification of green tea. Tea samples produced in three regions of Chongqing (southeastern Chongqing, northeastern Chongqing, and western Chongqing) for four months (from May to August 2020) were collected. Principal component analysis (PCA) was used to reduce data dimensionality and visualize the clustering of samples in different categories. Linear partial least squares-discriminant analysis (PLS-DA) and nonlinear support vector machine (SVM) algorithms were used to develop discriminant models. The PCA-SVM models based on the first four and first five principal components (PCs) achieved the best accuracies of 97.5% and 95% in the prediction set for geographical origin and processing month of green tea, respectively. This study demonstrated the feasibility of HSI in the identification of green tea species, providing a rapid and nondestructive method for the evaluation and control of green tea quality.

Biochemical methane potential prediction for mixed feedstocks of straw and manure in anaerobic co-digestion

To rapidly estimate the biochemical methane potential (BMP) of feedstocks, different multivariate regression models were established between BMP and the physicochemical indexes or near-infrared spectroscopy (NIRS). Mixed fermentation feedstocks of corn stover and livestock manure were rapidly detected BMP in anaerobic co-digestion (co-AD). The results showed that the predicted accuracy of NIRS model based on characteristic wavelengths selected by multiple competitive adaptive reweighted sampling outperformed all regression models based on the physicochemical indexes. For the NIRS regression model, coefficient of determination, root mean squares error, relative root mean squares error, mean relative error and residual predictive deviation of the validation set were 0.982, 6.599, 2.713%, 2.333% and 7.605. The results reveal that the predicted accuracy of NIRS model is very high, and meet the requirements of rapid prediction of BMP for co-AD feedstocks in practical biogas engineering.

M3GPSpectra: A novel approach integrating variable selection/construction and MLR modeling for quantitative spectral analysis

Quantitative analysis of the physical or chemical properties of various materials by using spectral analysis technology combined with chemometrics has become an important method in the field of analytical chemistry. This method aims to build a model relationship (called prediction model) between feature variables acquired by spectral sensors and components to be measured. Feature selection or transformation should be conducted to reduce the interference of irrelevant information on the prediction model because original spectral feature variables contain redundant information and massive noise. Most existing feature selection and transformation methods are single linear or nonlinear operations, which easily lead to the loss of feature information and affect the accuracy of subsequent prediction models. This research proposes a novel spectroscopic technology-oriented, quantitative analysis model construction strategy named M3GPSpectra. This tool uses genetic programming algorithm to select and reconstruct the original feature variables, evaluates the performance of selected and reconstructed variables by using multivariate regression model (MLR), and obtains the best feature combination and the final parameters of MLR through iterative learning. M3GPSpectra integrates feature selection, linear/nonlinear feature transformation, and subsequent model construction into a unified framework and thus easily realizes end-to-end parameter learning to significantly improve the accuracy of the prediction model. When applied to six types of datasets, M3GPSpectra obtains 19 prediction models, which are compared with those obtained by seven linear or non-linear popular methods. Experimental results show that M3GPSpectra obtains the best performance among the eight methods tested. Further investigation verifies that the proposed method is not sensitive to the size of the training samples. Hence, M3GPSpectra is a promising spectral quantitative analytical tool.

Rapid Biochemical Methane Potential Evaluation of Anaerobic Co-Digestion Feedstocks Based on Near Infrared Spectroscopy and Chemometrics

Biochemical methane potential (BMP) of anaerobic co-digestion (co-AD) feedstocks is an essential basis for optimizing ratios of materials. Given the time-consuming shortage of conventional BMP tests, a rapid estimated method was proposed for BMP of co-AD—with straw and feces as feedstocks—based on near infrared spectroscopy (NIRS) combined with chemometrics. Partial least squares with several variable selection algorithms were used for establishing calibration models. Variable selection methods were constructed by the genetic simulated annealing algorithm (GSA) combined with interval partial least squares (iPLS), synergy iPLS, backward iPLS, and competitive adaptive reweighted sampling (CARS), respectively. By comparing the modeling performances of characteristic wavelengths selected by different algorithms, it was found that the model constructed using 57 characteristic wavelengths selected by CARS-GSA had the best prediction accuracy. For the validation set, the determination coefficient, root mean square error and relative root mean square error of the CARS-GSA model were 0.984, 6.293 and 2.600, respectively. The result shows that the NIRS regression model—constructed with characteristic wavelengths, selected by CARS-GSA—can meet actual detection requirements. Based on a large number of samples collected, the method proposed in this study can realize the rapid and accurate determination of the BMP for co-AD raw materials in biogas engineering.

Rapid determination of lignocellulose in corn stover based on near-infrared reflectance spectroscopy and chemometrics methods

In this study, a rapid detection method based on near-infrared reflectance spectroscopy was proposed for measuring the contents of cellulose, hemicellulose and lignin in corn stover. In the basis of strategies of variable selection, feature extraction and nonlinear modeling, BiPLS-PCA-SVM was constructed using backward interval partial least squares combined with principal component analysis and support vector machine, which was used to improve the performance of spectral regression calibration model. For BiPLS-PCA-SVM model, the determination coefficients, root mean squared error and residual predictive deviation for the validation set were 0.906, 0.900% and 3.213 for cellulose; 0.987, 0.797% and 9.071 for hemicellulose; and 0.936, 0.264% and 4.024 for lignin, correspondingly. The results indicate that near-infrared reflectance spectroscopy combined with BiPLS-PCA-SVM can provide a reliable alternative strategy to detect contents of lignocellulosic components for pretreated corn stover in the anaerobic digestion process.