Uninformative Variable Elimination Research Articles

Rice Origin Tracing Technology Based on Fluorescence Spectroscopy and Stoichiometry

The origin of agricultural products is crucial to their quality and safety. This study explored the differences in chemical composition and structure of rice from different origins using fluorescence detection technology. These differences are mainly affected by climate, environment, geology and other factors. By identifying the fluorescence characteristic absorption peaks of the same rice seed varieties from different origins, and comparing them with known or standard samples, this study aims to authenticate rice, protect brands, and achieve traceability. The study selected the same variety of rice seed planted in different regions of Jilin Province in the same year as samples. Fluorescence spectroscopy was used to collect spectral data, which was preprocessed by normalization, smoothing, and wavelet transformation to remove noise, scattering, and burrs. The processed spectral data was used as input for the long short-term memory (LSTM) model. The study focused on the processing and analysis of rice spectra based on NZ-WT-processed data. To simplify the model, uninformative variable elimination (UVE) and successive projections algorithm (SPA) were used to screen the best wavelengths. These wavelengths were used as input for the support vector machine (SVM) prediction model to achieve efficient and accurate predictions. Within the fluorescence spectral range of 475–525 nm and 665–690 nm, absorption peaks of nicotinamide adenine dinucleotide (NADPH), riboflavin (B2), starch, and protein were observed. The origin tracing prediction model established using SVM exhibited stable performance with a classification accuracy of up to 99.5%.The experiment demonstrated that fluorescence spectroscopy technology has high discrimination accuracy in tracing the origin of rice, providing a new method for rapid identification of rice origin.

Rapid Measurement of Total Saponins, Mannitol, and Naringenin in Dendrobium officinale by Near-Infrared Spectroscopy and Chemometrics.

Dendrobium officinale has drawn increasing attention as a dual-use plant with herbal medicine and food applications. The efficient quality evaluation of D. officinale is essential to ensuring its nutritional and pharmaceutical value. Given that traditional analytical methods are generally time-consuming, expensive, and laborious, this study developed a rapid and efficient approach to assess the quality of D. officinale from different geographical origins by near-infrared (NIR) spectroscopy and chemometrics. Total saponins, mannitol, and naringenin were utilized as quality indicators. Two wavelength selection methods, namely, uninformative variable elimination and competitive adaptive reweighted sampling (CARS), were utilized to enhance the prediction accuracy of the quantification model. Moreover, multiple spectral pretreatment methods were applied for model optimization. Results indicated that the partial least squares (PLS) model constructed based on the wavelengths selected by CARS exhibited superior performance in predicting the contents of the quality indicators. The coefficient of determination (RP2) and root mean square error (RMSEP) in the independent test sets were 0.8949 and 0.1250 g kg-1 for total saponins, 0.9664 and 0.2192 g kg-1 for mannitol, and 0.8570 and 0.003159 g kg-1 for naringenin, respectively. This study revealed that NIR spectroscopy and the CARS-PLS model could be used as a rapid and accurate technique to evaluate the quality of D. officinale.

Design of printing ink spectral collection system and research on ink proportion prediction method

In spectral-based ink content measurement, improving the accuracy of the spectral collection system and reducing the algorithm complexity of the ink content measurement model are important aspects of research. This study designs a spectral collection system with an M-type Czerny–Turner optical path structure, uses a third-order polynomial fitting method for calibration analysis, collects the spectral of the standard color card and denoises it through Savitzky–Golay convolution smoothing, and establishes a functional relationship between the spectral and the logarithm of ink content. Uninformative Variable Elimination (UVE) and Competitive Adaptive Reweighting Sampling are used to perform comparative analysis on feature wavelength extraction, and a prediction model for printing ink content is established. Experimental results show that the R2 of ink C, ink M, and ink Y are 0.9981, 0.9975, and 0.9892, respectively, and the root mean square error is 0.0134, 0.0153, and 0.0317, respectively, showing good performance in ink prediction accuracy.

Simultaneous Rapid Detection of Multiple Physicochemical Properties of Jet Fuel Using Near-Infrared Spectroscopy.

Jet fuel is the primary fuel used in the aviation industry, and its quality has a direct impact on the safety and operational efficiency of aircraft. The accurate quantitative detection and analysis of various physicochemical property indicators are important for improving and ensuring the quality of jet fuel in the domestic market. This study used near-infrared (NIR) spectroscopy to establish a suitable model for the simultaneous and rapid detection of multiple physicochemical properties in jet fuel. Using more than 40 different sources of jet fuel, a rapid detection model was established by optimizing the spectral processing methods. The measurement models were separately built using the partial least-squares (PLS) and orthogonal PLS algorithms, and the model parameters were optimized. The results show that after the Savitzky-Golay second derivative preprocessing, the PLS model built using the feature spectra selected by the uninformative variable elimination wavelength algorithm achieved the best measurement performance. Compared with the PLS model without preprocessing, the range of the resulting accuracy improvement was at least 15.01%. Under the optimal model parameters, the calibration set regression coefficient (Rc2) of the 11 jet fuel property index models ranged from 0.9102 to 0.9763, with the root-mean-square error of calibration values up to 0.8468 °C (for flash points). The regression coefficient (Rp2) of the validation set ranged from 0.8239 to 0.9557, with the root-mean-square error of prediction values up to 1.1354 °C (for flash points). The ratios of prediction to deviation (RPD) values were all in the range of 1.9-3.0, indicating high accuracy and reliability of the model. The rapid NIR analysis method established in this study enables the simultaneous and rapid detection of multiple physicochemical properties of jet fuel, thereby providing effective technical support for ensuring the quality of jet fuel in the market.

Development of near-infrared spectroscopy calibration model and monitoring software: For monitoring hexamethylenetetramine concentration in hexamethylenetetramine–acetic acid solution

In recent years, near-infrared spectroscopy has been increasingly used in chemical engineering processes. However, accurately and in real-time monitoring concentrations remains highly challenging. This study presents a method for accurately monitoring the hexamethylenetetramine concentration in hexamethylenetetramine-acetic acid solution by integrating near-infrared spectroscopy with software techniques. Compared to the methods of zero mean standardization (Z-score), standard normal variate (SNV), first derivative (D1), second derivative (D2), and vector normalization (VN), the Savitzky-Golay first derivative (SGFD) method effectively addresses issues related to baseline drift, noise, and other interferences. This results in clearer identification of characteristic absorption peaks of different components and a significant improvement in the accuracy of the analysis. The study combined the uninformative variable elimination (UVE), differential evolution (DE), and moth flame optimization (MFO) algorithms to optimize a feature wavenumber selection method, which reduced the number of variables by 78.74 %, and the values of root mean square error of calibration set (RMSEC) and root mean square error of prediction (RMSEP) by 18.86 % and 30.38 %, respectively. Additionally, the analytical method was integrated with real-time monitoring software, enabling the monitoring of hexamethylenetetramine concentration. This integration supports the improvement of the quality and safety of the manufacturing process. The results demonstrate the potential application of near-infrared spectroscopy in real-time monitoring and automated control of chemical processes in the chemical industry.

Accurate identification of methanol and ethanol gasoline types and rapid detection of the alcohol content using effective chemical information

Methanol and ethanol gasoline are two emerging clean energy sources with different characteristics. To achieve the qualitative identification and quantitative analysis of the alcohols present in methanol and ethanol gasoline, effective chemical information (ECI) models based on the characteristic spectral bands of the near-infrared (NIR) spectra of the methanol and ethanol molecules were developed using the partial least squares discriminant analysis (PLS-DA) and partial least squares (PLS) algorithms. The ECI model was further compared with models built from the full wavenumber (Full) spectra, variable importance in projection (VIP) spectra, and Monte Carlo uninformative variable elimination (MC-UVE) spectra to determine the predictive performance of ECI model. Among the various qualitative identification models, it was found that the ECI-PLS-DA model, which is built using the differences in molecular chemical information between methanol and ethanol, exhibited sensitivity, specificity and accuracy values of 100%. The ECI-PLS-DA model accurately identified methanol gasoline and ethanol gasoline with different contents. In the quantitative analysis model for methanol gasoline, the methanol gasoline and ethanol gasoline ECI-PLS models exhibited the smallest root mean squared error of predictions (RMSEP) of 0.18 and 0.21% (v/v), respectively, compared to the other models. Meanwhile, the F-test and T-test results revealed that the NIR method employing the ECI-PLS model showed no significant difference compared to the standard method. Compared with other spectral models examined herein, the ECI model demonstrated the highest recognition success and determination accuracy. This study therefore established a highly accurate and rapid determination model for the qualitative identification and quantitative analysis based on chemical structures. It is expected that this model could be extended to the NIR analysis of other physicochemical properties of fuel.

Unsupervised Clustering-Assisted Method for Consensual Quantitative Analysis of Methanol-Gasoline Blends by Raman Spectroscopy.

Methanol-gasoline blends have emerged as a promising and environmentally friendly bio-fuel option, garnering widespread attention and promotion globally. The methanol content within these blends significantly influences their quality and combustion performance. This study explores the qualitative and qualitative analysis of methanol-gasoline blends using Raman spectroscopy coupled with machine learning methods. Experimentally, methanol-gasoline blends with varying methanol concentrations were artificially configured, commencing with initial market samples. For qualitative analysis, the partial least squares discriminant analysis (PLS-DA) model was employed to classify the categories of blends, demonstrating high prediction performance with an accuracy of nearly 100% classification. For the quantitative analysis, a consensus model was proposed to accurately predict the methanol content. It integrates member models developed on clustered variables, using the unsupervised clustering method of the self-organizing mapping neural network (SOM) to accomplish the regression prediction. The performance of this consensus model was systemically compared to that of the PLS model and uninformative variable elimination (UVE)-PLS model. Results revealed that the unsupervised consensus model outperformed other models in predicting the methanol content across various types of methanol gasoline blends. The correlation coefficients for prediction sets consistently exceeded 0.98. Consequently, Raman spectroscopy emerges as a suitable choice for both qualitative and quantitative analysis of methanol-gasoline blend quality. This study anticipates an increasing role for Raman spectroscopy in analysis of fuel composition.

Rapid Quality Assessment of Polygoni Multiflori Radix Based on Near-Infrared Spectroscopy

The precise and prompt determination of quality control indicators such as moisture, stilbene glycosides, and anthraquinone glycosides is crucial in assessing the quality of Polygoni Multiflori Radix. Near-infrared spectroscopy is a nondestructive analytical technique that offers a more desirable approach than traditional methods for assessing content levels. In this study, various spectral preprocessing techniques were used to preprocess the raw spectral data. The spectral data were correlated with the determination of three-component contents using the partial least squares regression (PLSR) method. Then different algorithms, such as competitive adaptive weighted sampling (CARS), Monte Carlo uninformative variable elimination (MCUVE), and random frog hopping (RF), were used for model simplification and feature selection. The data suggest that the first-order deconvolution derivative (1st Dev.) processing of the spectral data is superior to other methods in all three model evaluation metrics. The PLSR model for moisture, stilbene glycosides, and anthraquinone glycosides produced the calibration coefficient of determination (R2C) of 0.82, 0.52, and 0.58, the root mean square error of cross validation (RMSECV) of 0.91%, 0.77%, and 0.69%, the prediction coefficient of determination (R2P) of 0.72, 0.28, and 0.54, the root mean square error of prediction (RMSEP) of 0.65%, 0.81%, and 0.75%, and relative percentage differences (RPDs) of 1.7, 1.0, and 0.8. After optimizing the model using CARS, R2C increased by 0.15%, 0.41%, and 0.34%, RMSECV decreased by 0.53%, 0.32%, and 0.24%, R2P increased by 0.21%, 0.63%, and 0.35%, RMSEP decreased by 0.36%, 0.41%, and 0.31%, and RPD increased by 1.1, 0.9, and 0.6, significantly improving the predictive capacity of the model. This research provides a feasible method for rapid compliance testing of Polygoni Multiflori Radix. To further improve the model’s performance and applicability, it is necessary to continuously expand the sample set with different varieties and locations for wide variation.

Deoxynivalenol Detection beyond the Limit in Wheat Flour Based on the Fluorescence Hyperspectral Imaging Technique.

Deoxynivalenol (DON) is a harmful fungal toxin, and its contamination in wheat flour poses a food safety concern globally. This study proposes the combination of fluorescence hyperspectral imaging (FHSI) and qualitative discrimination methods for the detection of excessive DON content in wheat flour. Wheat flour samples were prepared with varying DON concentrations through the addition of trace amounts of DON using the wet mixing method for fluorescence hyperspectral image collection. SG smoothing and normalization algorithms were applied for original spectra preprocessing. Feature band selection was carried out by applying the successive projection algorithm (SPA), uninformative variable elimination (UVE), competitive adaptive reweighted sampling (CARS), and the random frog algorithm on the fluorescence spectrum. Random forest (RF) and support vector machine (SVM) classification models were utilized to identify wheat flour samples with DON concentrations higher than 1 mg/kg. The results indicate that the SG-CARS-RF and SG-CARS-SVM models showed better performance than other models, achieving the highest recall rate of 98.95% and the highest accuracy of 97.78%, respectively. Additionally, the ROC curves demonstrated higher robustness on the RF algorithm. Deep learning algorithms were also applied to identify the samples that exceeded safety standards, and the convolutional neural network (CNN) model achieved a recognition accuracy rate of 97.78% for the test set. In conclusion, this study demonstrates the feasibility and potential of the FHSI technique in detecting DON infection in wheat flour.

Nondestructive Determination of Epicarp Hardness of Passion Fruit Using Near-Infrared Spectroscopy during Storage.

The hardness of passion fruit is a critical feature to consider when determining maturity during post-harvest storage. The capacity of near-infrared diffuse reflectance spectroscopy (NIRS) for non-destructive detection of outer and inner hardness of passion fruit epicarp was investigated in this work. The passion fruits' spectra were obtained using a near-infrared spectrometer with a wavelength range of 10,000-4000 cm-1. The hardness of passion fruit's outer epicarp (F1) and inner epicarp (F2) was then measured using a texture analyzer. Moving average (MA) and mean-centering (MC) techniques were used to preprocess the collected spectral data. Competitive adaptive reweighted sampling (CARS), successive projection algorithm (SPA), and uninformative variable elimination (UVE) were used to pick feature wavelengths. Grid-search-optimized random forest (Grids-RF) models and genetic-algorithm-optimized support vector regression (GA-SVR) models were created as part of the modeling process. After MC preprocessing and CARS selection, MC-CARS-Grids-RF model with 7 feature wavelengths had the greatest prediction ability for F1. The mean square error of prediction set (RMSEP) was 0.166 gN. Similarly, following MA preprocessing, the MA-Grids-RF model displayed the greatest predictive performance for F2, with an RMSEP of 0.101 gN. When compared to models produced using the original spectra, the R2P for models formed after preprocessing and wavelength selection improved. The findings showed that near-infrared spectroscopy may predict the hardness of passion fruit epicarp, which can be used to identify quality during post-harvest storage.

Application of hyperspectral imaging and colorimetric sensor array coupled with multivariate analysis for quality detection during salted duck eggs processing

AbstractIn order to realize the evaluation of salted duck eggs quality, this study constructed salts‐sensitive colorimetric sensor array (CSA) based on silver nanoprisms (AgNPRs) etching principle and utilized hyperspectral imaging (HSI) system to capture the reflectance spectral of CSA and section of salted duck eggs. To minimize the noise in spectral signals, three preprocessing methods including Savitzky–Golay Smoothing (SG), standard normal variate (SNV) transformation, and multiplicative scatter correction were used. The competitive adaptive reweighted sampling and uninformative variables elimination were used to select the optimal wavelengths, and then partial least squares regression was used to predict moisture, lipid, and salt content (SC). Results demonstrated coefficients of determination of 0.8200, 0.8969, 0.8709, and 0.9239 for protein moisture, yolk moisture, yolk lipid, and protein SC. Visualization distribution maps successfully display the distribution of moisture and lipid content during salting periods. The results indicated that the CSA based on AgNPRs etching principle can be an effective tool in quantitative measurement of salt and HSI combining CSA can be applied to evaluate the quality of salted duck eggs comprehensively. The study offers a rapid detection method for the production and processing of salted duck eggs in the industry.Practical applicationA novel salts‐sensitive CSA was constructed to evaluate quality of salted duck eggs based on AgNPRs etching. This study designed and developed a portable sample collection device using 3D printing technology to acquire the information of samples and CSA at the same time. This study combined with HSI system and colorimetric sensor technology to evaluate the quality of salted duck eggs during the salting process.

Quantitative analysis of reducing sugars in solid‐state fermentation based on NIR combined with GA‐AdaBoost

SummaryThis study focuses primarily on the real‐time monitoring of reducing sugar levels throughout solid‐state ethanol fermentation. A predictive model for accurate quantification of reducing sugar levels will be formulated through a comprehensive analysis of near‐infrared spectra across different fermentation stages. Building a predictive model using Adaptive Boosting (AdaBoost), The spectral data was pre‐processed using the Standard Normal Variate (SNV) method, followed by feature band selection using the Genetic Algorithm (GA), resulting in the identification of 181 characteristic wavelengths. The resulting GA‐AdaBoost model has excellent performance metrics: of 0.989, RMSEC of 0.393, of 0.978, RMSEP of 0.640, RPD of 6.651 and RER of 0.142. Comparative evaluations against Competitive Adaptive Reweighted Sampling (CARS), Successive Projection Algorithm (SPA), Uninformative Variable Elimination (UVE) and Principal Component Analysis (PCA) consistently underscore the superior efficacy of the GA‐AdaBoost model. These results robustly confirm its reliability in accurately quantifying reducing sugars in solid‐state fermentation processes.

Application of Near-Infrared Spectroscopy to Rapidly Classify the Chinese Quince Fruits from Different Habitats

The ecological habitats of Chinese quince (Chaenomeles speciosa Nakai) fruits affect their phenotype. Currently, limited or no rapid method exists for classifying Chinese quince fruit from different ecosystems. This study developed a partial least squares discriminant analysis (PLS-DA) classification model to effectively and nondestructively classify 663 Chinese quince fruit samples from six environments in 2020. PLS-DA models and other variable selection approaches were used in this study. The near-infrared spectroscopy (NIRs) absorption spectra of raw Chinese quince fruit samples from six habitats showed a similar shape. The spectra of each environment showed little variance. The raw fruit spectra varied significantly among habitat categories after the first derivative preprocessing phase. The uninformative variable elimination (UVE) variable selection approach had greater calibration and validation set specificity of 0.93 and 0.98. This study found the best classification specificity using the UVE variable selection approach compared to other methods including the PLS-DA model without variable selection. The UVE approach improved Yunnan habitat categorization specificity from 86% to 88% when integrated with PLS-DA. Additionally, the validation set for quinces originating from Anhui, Chongqing, Hubei, Shandong, and Zhejiang achieved an ideal classification score of 100%. The findings of the study indicated that PLS-DA can serve as an alternative approach for classifying the habitats of Chinese quince fruits. When used in conjunction with other methods, this technique can assist researchers, scientists, and industry professionals in identifying the main factors responsible for significant variations in the habitats, composition, and quality of Chinese quince fruits.

Origin traceability and adulteration detection of soybean using near infrared hyperspectral imaging

AbstractStable isotopes, multi‐elements, metabolic profiles, and integrated spectroscopic fingerprints are priority options for food geographical origin traceability. However, til now, it is still hard to detect adteration with the same one from other geographic origins, which is harder than geographical origin traceability. In this study, partial least square discriminant analysis was employed to build a classification model to discriminate the domestic and imported soybeans after variable selection by uninformative variable elimination using near infrared hyperspectral imaging. As a result, this model could completely discriminate domestic and imported soybeans. Moreover, the developed model was used to detect the adulterated domestic soybean was adulterated with 13.3%, 20.0%, 26.7%, and 33.3% of imported soybean. When the skewness value was less than 0.76 and kurtosis value was less than 1.57 of a sample, the sample was considered as the adulterated. The results indicated that the domestic soybeans adulterated with 20.0%, 26.7%, and 33.3% of imported soybeans were successfully identified. This method could not only identify origin traceability but also detect adulteration of soybeans, which will be beneficial to guarantee the quality and safety of soybean.

Establishment of a general prediction model for protein content in various varieties and colors of peas using visible-near-infrared spectroscopy

Non-destructive and swift detection of protein content in peas holds significant importance for screening superior germplasm resources. This study aimed to construct a general detection model of pea protein content. Spectral detection data of pea proteins were analyzed across three bands: 400 ∼ 982 nm, 982 ∼ 2520 nm, and 400 ∼ 2520 nm. The spectral data were preprocessed by 16 different methods. The successive projections algorithm (SPA) and uninformative variables elimination (UVE) algorithms were used to screen characteristic wavelengths. Pea protein content prediction models were established by using partial least squares regression (PLSR) and extreme learning machine (ELM). Following analysis, the best modeling band was 400 ∼ 982 nm. The optimal prediction model was multiplicative scatter correction (MSC)-SPA-ELM. The correlation coefficients for the calibration and validation sets were 0.8851 and 0.8595, respectively, with root mean square errors of 0.8933 g/100 g and 1.0565 g/100 g.The ratio of error range (RER) was 9.23. These results suggest that this study offers a viable solution for non-destructive and rapid pea protein detection across diverse regions, varieties, and colors.

Determination of the Acid and Peroxide Values of Vegetable Oils by Raman Spectroscopy with Competitive Adaptive Reweighted Sampling (CARS) and Back Propagation Neural Network (BPNN)

Acid and peroxide values are important indicators to evaluate the hydrolysis or oxidation level of vegetable oil. In order to achieve rapid determination of these parameters by Raman spectroscopy, rapeseed and soybean oil were analyzed, and a competitive adaptive reweighted sampling-back propagation neural network (CARS-BPNN) method was developed. First, the Raman spectrum was preprocessed to reduce noise using moving window smoothing. Next, competitive adaptive reweighted sampling (CARS), successive projections algorithm (SPA), and uninformative variable elimination (UVE) was applied to extract the Raman characteristic information that indicated the acid and peroxide values. Based on the extracted characteristic information, back propagation neural network (BPNN), partial least squares regression (PLSR), and support vector machine (SVM) prediction models were established and compared. The results showed that the best accuracy was obtained using CARS-BPNN. For rapeseed oil, the coefficients of determination (R2) of acid value and peroxide value were 0.9207 and 0.9772. The root mean square error (RMSE) values were 0.0909 and 0.0022, respectively. For soybean oil, the coefficients of determination (R 2) of acid and peroxide values were 0.9381 and 0.9716. The RMSE values were 0.1250 and 0.0027, respectively. A good correlation coefficient and a small RMSE value were obtained. The results indicate that Raman spectroscopy combined with CARS-BPNN accurately evaluated the acid and peroxide values, which provides a reference for the optimization of the Raman spectrum and creates a new approach for the evaluation of vegetable oil quality.

Optimization of Fourier transform near-infrared spectroscopy model in determining saponin compounds of Panax notoginseng roots

As a traditional Chinese medicine, Panax notoginseng (Burk.) F.H.Chen (P. notoginseng) is abundant in chemical compounds, particularly the high content of saponin compounds, which have been extensively implemented in clinical treatment. The traditional chemical methods have drawbacks of destroying samples and taking a long time to analyze the saponin compounds content. In this study, we investigated the viability of employing Fourier transform near infrared spectroscopy (FT-NIR) to assess the saponin compounds content of P. notoginseng rapidly. The partial least squares regression (PLSR) prediction model was established based on spectral information from 252 samples. The effects of various variable selection methods, including variable importance in projection (VIP), competitive adaptive reweighted sampling (CARS), uninformative variables elimination (UVE), and correlation coefficients (Correlation) on the model performance, were compared. One examined variable selection algorithm that stood out was the correlation coefficient method. The Correlation-PLSR model’ calibration and prediction sets had a high coefficient of determination (Rc2: 0.966-0.989; Rp2: 0.968-0.999) and low root mean square error (RMSEC: 1.293-5.984; RMSEP: 0.291-1.810). It was indicated it can rapidly predict saponin compounds in P. notoginseng. This study offers a rapid and reliable quantitative method for P. notoginseng quality control.

Rapid determination of two illegal additives in wheat flour by near-infrared spectroscopy and different key wavelength selection algorithms

Rapid and accurate testing of wheat flour for illegal additives is important for maintaining food safety. In this study, a combination of NIR spectroscopy and partial least squares regression (PLSR) was used to predict the levels of two illegal additives (talc powder and benzoyl peroxide) in wheat flour. Different critical wavelength selection algorithms (competitive adaptive reweighted sampling, CARS; random frog, RF; Monte Carlo uninformative variable elimination, MCUVE; CARS-RF; CARS-MCUVE) were used to filter key wavelengths. In our study, R2P for talc and benzoyl peroxide in the full-wavelength PLSR model was 0.995 and 0.964, respectively. The CARS-RF and RF were screened for 10 effective wavelengths. In wheat flour containing talc, the R2P and RPD of the model using the CARS-RF were 0.995 and 14.796, respectively. In wheat flour containing benzoyl peroxide, the R2P and RPD of the model using the RF were 0.923 and 3.697, respectively. CARS-RF and RF algorithms had great potential to accurately predict illegal additives.

Regional Inversion of Soil Heavy Metal Cr Content in Agricultural Land Using Zhuhai-1 Hyperspectral Images.

With the development of hyperspectral imaging technology, the potential for utilizing hyperspectral images to accurately estimate heavy metal concentrations in regional soil has emerged. Currently, soil heavy metal inversion based on laboratory hyperspectral data has demonstrated a commendable level of accuracy. However, satellite images are susceptible to environmental factors such as atmospheric and soil background, presenting a significant challenge in the accurate estimation of soil heavy metal concentrations. In this study, typical chromium (Cr)-contaminated agricultural land in Shaoguan City, Guangdong Province, China, was taken as the study area. Soil sample collection, Cr content determination, laboratory spectral measurements, and hyperspectral satellite image collection were carried out simultaneously. The Zhuhai-1 hyperspectral satellite image spectra were corrected to match laboratory spectra using the direct standardization (DS) algorithm. Then, the corrected spectra were integrated into an optimal model based on laboratory spectral data and sample Cr content data for regional inversion of soil heavy metal Cr content in agricultural land. The results indicated that the combination of standard normal variate (SNV)+ uninformative variable elimination (UVE)+ support vector regression (SVR) model performed best with laboratory spectral data, achieving a high accuracy with an R2 of 0.97, RMSE of 5.87, MAE of 4.72, and RPD of 4.04. The DS algorithm effectively transformed satellite hyperspectral image data into spectra resembling laboratory measurements, mitigating the impact of environmental factors. Therefore, it can be applied for regional inversion of soil heavy metal content. Overall, the study area exhibited a low-risk level of Cr content in the soil, with the majority of Cr content values falling within the range of 36.21-76.23 mg/kg. Higher concentrations were primarily observed in the southeastern part of the study area. This study can provide useful exploration for the promotion and application of Zhuhai-1 image data in the regional inversion of soil heavy metals.

Hyperspectral Prediction Model of Nitrogen Content in Citrus Leaves Based on the CEEMDAN–SR Algorithm

Nitrogen content is one of the essential elements in citrus leaves (CL), and many studies have been conducted to determine the nutrient content in CL using hyperspectral technology. To address the key problem that the conventional spectral data-denoising algorithms directly discard high-frequency signals, resulting in missing effective signals, this study proposes a denoising preprocessing algorithm, complete ensemble empirical mode decomposition with adaptive noise joint sparse representation (CEEMDAN–SR), for CL hyperspectral data. For this purpose, 225 sets of fresh CL were collected at the Institute of Fruit Tree Research of the Guangdong Academy of Agricultural Sciences, to measure their elemental nitrogen content and the corresponding hyperspectral data. First, the spectral data were preprocessed using CEEMDAN–SR, Stein’s unbiased risk estimate and the linear expansion of thresholds (SURE–LET), sparse representation (SR), Savitzky–Golay (SG), and the first derivative (FD). Second, feature extraction was carried out using principal component analysis (PCA), uninformative variables elimination (UVE), and the competitive adaptive re-weighted sampling (CARS) algorithm. Finally, partial least squares regression (PLSR), support vector regression (SVR), random forest (RF), and Gaussian process regression (GPR) were used to construct a CL nitrogen prediction model. The results showed that most of the prediction models preprocessed using the CEEMDAN–SR algorithm had better accuracy and robustness. The prediction models based on CEEMDAN–SR preprocessing, PCA feature extraction, and GPR modeling had an R2 of 0.944, NRMSE of 0.057, and RPD of 4.219. The study showed that the CEEMDAN–SR algorithm can be effectively used to denoise CL hyperspectral data and reduce the loss of effective information. The prediction model using the CEEMDAN–SR+PCA+GPR algorithm could accurately obtain the nitrogen content of CL and provide a reference for the accurate fertilization of citrus trees.