Variable Iterative Space Shrinkage Approach Research Articles

Quantitative analysis of wool and cashmere fiber mixtures using NIR spectroscopy

Abstract The quantitative determination of wool and cashmere mixed fiber is an indispensable quality control link in the textile industry, crucial for improving international trade status, ensuring product quality, and safeguarding consumer rights. Therefore, the goal of this study is to develop a reliable method for estimating fiber contents in wool–cashmere blends based on near-infrared (NIR) spectroscopy. A total of 210 mixed samples of 21 different proportions of cashmere and wool are prepared in the experiment, and data are collected in the NIR spectral band of 1,000–2,500 nm. Convolution Savitzky–Golay (S–G) combined with the second-order derivative is then used for spectral preprocessing. The variable iterative space shrinkage approach (VISSA) optimizes the characteristic wavelengths, and 339 wavelength points are selected. The prediction model of the least squares support vector machine (LSSVM) is established by particle swarm optimization (PSO), fast positioning, and analysis of key information related to the target in complex spectral data. Finally, the training set and the prediction set are divided according to the ratio of 8 : 2. Experiments show that in terms of modeling and prediction, the PSO-LSSVM model based on the wavelength selected by VISSA has a prediction determination coefficient R-squared of 0.9821, a prediction root mean square error of 1.1263, and an mean absolute error of 0.6527. The hybrid modeling method of VISSA, PSO, and LSSVM based on NIR spectroscopy (VISSA–PSO–LSSVM) can provide a more accurate and stable method for the non-destructive detection of cashmere and wool blended fiber content.

Rapid qualitative and quantitative detection for adulteration of Atractylodis Rhizoma using hyperspectral imaging combined with chemometric methods

In the field of traditional Chinese medicine, Atractylodis Rhizoma (AR) is commonly used for various diseases due to its excellent ability to dry dampness and strengthen the spleen, especially popular in East Asia. The aim of this study is to proposed Hyperspectral Imaging (HSI) in combination with chemometric methods for the rapid qualitative and quantitative detection of AR adulteration with other types of powder. Partial Least Squares Discriminant Analysis (PLS-DA) was used to construct the classification models the best, with the First-order Derivative (F-D) preprocessing method. The accuracy values of the test sets for classification models were above 99%. Furthermore, Partial Least Squares Regression (PLSR), Random Forest Regression (RFR), and BP Neural Network (BPNN) were used to quantitatively analyze the adulteration level. On the whole, the BPNN model has a relatively stable effect. The R-square (R2) values of different models were all greater than 0.97, the Root Mean Square Error (RMSE) values were all less than 0.0300, and the Relative Percentage Difference (RPD) values were over 6.00. After applying three characteristic wavelength selection algorithms, namely Iterative Retained Information Variable (IRIV), Successive Projections Algorithm (SPA), and Variable Iterative Space Shrinkage Approach (VISSA) algorithms, the classification accuracy values remained over 99.00% while the quantification models’ RPD values were over 4.00. These results demonstrate the reliability of using hyperspectral imaging combined with chemometrics methods for the adulteration problems in AR.

A Sensitive SERS Sensor Combined with Intelligent Variable Selection Models for Detecting Chlorpyrifos Residue in Tea.

Chlorpyrifos is one of the most widely used broad-spectrum insecticides in agriculture. Given its potential toxicity and residue in food (e.g., tea), establishing a rapid and reliable method for the determination of chlorpyrifos residue is crucial. In this study, a strategy combining surface-enhanced Raman spectroscopy (SERS) and intelligent variable selection models for detecting chlorpyrifos residue in tea was established. First, gold nanostars were fabricated as a SERS sensor for measuring the SERS spectra. Second, the raw SERS spectra were preprocessed to facilitate the quantitative analysis. Third, a partial least squares model and four outstanding intelligent variable selection models, Monte Carlo-based uninformative variable elimination, competitive adaptive reweighted sampling, iteratively retaining informative variables, and variable iterative space shrinkage approach, were developed for detecting chlorpyrifos residue in a comparative study. The repeatability and reproducibility tests demonstrated the excellent stability of the proposed strategy. Furthermore, the sensitivity of the proposed strategy was assessed by estimating limit of detection values of the various models. Finally, two-tailed paired t-tests confirmed that the accuracy of the proposed strategy was equivalent to that of gas chromatography-mass spectrometry. Hence, the proposed method provides a promising strategy for detecting chlorpyrifos residue in tea.

Discrimination of tea seed oil adulteration based on near-infrared spectroscopy and combined preprocessing method

Near-infrared spectroscopy and chemometrics was used to qualitatively distinguish the types of adulterated oils in binary adulteration of tea seed oil in this study. To address the limitations of a single preprocessing method, nine preprocessing methods from four categories were combined, and the impact of preprocessing method order on model accuracy was assessed. Additionally, variable iterative space shrinkage approach (VISSA), interval combinatorial optimization (ICO), and uninformative variables elimination (UVE) were used to screen characteristic wavelengths. Subsequently, a discriminative model for tea seed oil adulteration was constructed using two strategies. The results indicate that the order of preprocessing methods significantly influences model accuracy, and combining preprocessing methods can effectively enhance model accuracy. All three characteristic wavelength selection methods effectively screened characteristic variables. Both two strategies demonstrate good discriminant capabilities for binary adulteration in tea seed oil. In strategy 1, identification accuracies for the calibration, prediction and external datasets are 98.67 %, 100 % and 94.44 %, respectively. In strategy 2, identification accuracies for the calibration, prediction and external datasets are 100 %, 98 % and 94.44 %, respectively. Therefore, integrating NIRS with combined preprocessing and variable screening can effectively discern the types of adulterated oils in tea seed oil, serving as a potent detection tool.

Hyperspectral and Fluorescence Imaging Approaches for Nondestructive Detection of Rice Chlorophyll.

Estimating and monitoring chlorophyll content is a critical step in crop spectral image analysis. The quick, non-destructive assessment of chlorophyll content in rice leaves can optimize nitrogen fertilization, benefit the environment and economy, and improve rice production management and quality. In this research, spectral analysis of rice leaves is performed using hyperspectral and fluorescence spectroscopy for the detection of chlorophyll content in rice leaves. This study generated ninety experimental spectral datasets by collecting rice leaf samples from a farm in Sichuan Province, China. By implementing a feature extraction algorithm, this study compresses redundant spectral bands and subsequently constructs machine learning models to reveal latent correlations among the extracted features. The prediction capabilities of six feature extraction methods and four machine learning algorithms in two types of spectral data are examined, and an accurate method of predicting chlorophyll concentration in rice leaves was devised. The IVSO-IVISSA (Iteratively Variable Subset Optimization-Interval Variable Iterative Space Shrinkage Approach) quadratic feature combination approach, based on fluorescence spectrum data, has the best prediction performance among the CNN+LSTM (Convolutional Neural Network Long Short-Term Memory) algorithms, with corresponding RMSE-Train (Root Mean Squared Error), RMSE-Test, and RPD (Ratio of standard deviation of the validation set to standard error of prediction) indexes of 0.26, 0.29, and 2.64, respectively. We demonstrated in this study that hyperspectral and fluorescence spectroscopy, when analyzed with feature extraction and machine learning methods, provide a new avenue for rapid and non-destructive crop health monitoring, which is critical to the advancement of smart and precision agriculture.

Rapid detection of ash content in black tea using a homemade miniature near-infrared spectroscopy

Ash is a testing index with both health inspection value and quality decision value, and it is an essential detection item in the import and export trade of tea. To realize the rapid and effective quantitative analysis of ash content in tea, this study proposed the use of a homemade miniature near-infrared (NIR) spectroscopy combined with multivariate analysis for the rapid detection of ash content in black tea. First, NIR data of black tea samples from different countries were acquired and optimized by the spectral preprocessing method. Then, the optimized pre-processed spectral data were used as features, and four feature wavelength selection algorithms, such as competitive adaptive reweighted sampling, iteratively retaining informative variables (IRIV), variable combination population analysis (VCPA)-IRIV, and interval variable iterative space shrinkage approach (IVISSA), were utilized to optimize the feature spectra. Finally, the support vector machine regression (SVR) algorithm was employed to construct the quantitative models of ash content in black tea by combining the optimal wavelengths obtained from the four feature selection methods mentioned above. The experimental results showed that the IVISSA-SVR model had the best performance, with correlation coefficient (Rp), root mean square errors of prediction (RMSEP), and relative prediction deviation (RPD) of 0.9546, 0.0192, and 5.59 for the prediction set, respectively. The results demonstrate that a miniature NIR sensing system combined with chemometrics as an effective analytical tool can realize the rapid detection of ash content in black tea.

Quantitative detection of zearalenone in wheat using intervals selection coupled to near-infrared spectroscopy

Wheat is a globally cultivated cereal crop. However, it is susceptible to Fusarium head blight, which leads to the production of zearalenone (ZEN), significantly impacting the value of wheat. This paper proposed a novel technique for quantitatively detecting ZEN in wheat using near-infrared spectroscopy. The absorbance of wheat flour samples with varying degrees of mildew were collected using a spectral detection system based on the spectrometer. And this paper introduced two methods for feature intervals selection, interval combination optimization (ICO) and interval variable iterative space shrinkage approach (iVISSA), combined with partial least squares (PLS) and support vector machine (SVM) established a regression model based on feature intervals selection for quantitatively detecting ZEN in wheat, and then compared the performance of them. The research results showed that the SVM model outperformed the PLS model in terms of detection performance. Additionally, the SVM model detecting the level of ZEN in wheat based on the feature intervals selected by ICO achieved the best generalization performance, with a root mean square error of prediction (RMSEP) of 31.4148 μg·kg−1, a coefficient of determination of prediction (RP2) of 0.9434, and a relative percent deviation (RPD) of 4.2768. These results demonstrated that the SVM regression model, established by using ICO for selecting the near-infrared spectral intervals, can accurately detect ZEN in wheat. Near-infrared spectroscopy combined with appropriate data analysis methods, which can be used as a simple and fast tool for quantitative detection of mycotoxins of wheat and other grains.

Nondestructive testing and visualization of compound heavy metals in lettuce leaves using fluorescence hyperspectral imaging

This study evaluated the feasibility of nondestructive testing and visualization of compound heavy metals (cadmium and lead) in lettuce leaves using fluorescence hyperspectral imaging. In addition, a method involving wavelet transform and stepwise regression (WT-SR) was proposed to perform dimensionality reduction of fluorescence spectral data. Fluorescent hyperspectral image acquisition and mathematical analysis were carried out on lettuce leaf samples processed with different compound heavy metal concentrations. The entire lettuce leaf sample was selected as a region of interest (ROI). Savitzky-Golay (SG) algorithm, multivariate scatter correction (MSC), standard normalized variable (SNV), first derivative (1st Der) and second derivative (2nd Der) were used to preprocess the ROI fluorescence spectra. Further, the successive projections algorithm (SPA), the competitive adaptive reweighted sampling (CARS), the iteratively retaining informative variables (IRIV) and variable iterative space shrinkage approach (VISSA), and the wavelet transform combined with stepwise regression (WT-SR) were used to reduce the dimension of spectral data. Finally, the multiple linear regression (MLR) algorithm was used to build the compound heavy metal content detection models. The results showed that the MLR models based on the feature data obtained by 1st Der-WT-SR achieved reasonable performance with Rp2 of 0.7905, RMSEP of 0.0269 mg/kg and RPD of 2.477 for Cd content under wavelet fifth layer decomposition, and with Rp2 of 0.8965, RMSEP of 0.0096 mg/kg and RPD of 3.211 for Pb content under wavelet first layer decomposition. The distribution maps of cadmium and lead contents in lettuce leaves were established using the optimal prediction models. The results further confirmed the great potential of fluorescence hyperspectral technology combined with optimization algorithm for the detection of compound heavy metals.

Fusion of Spectral and Textural Data of Hyperspectral Imaging for Glycine Content Prediction in Beef Using SFCN Algorithms

Glycine, the simplest free amino acid. It is one of the important factors affecting the flavor of beef. In this study, a fast and non-destructive method combining near-infrared hyperspectral (900–1700 nm) and textural data was first proposed to determine the content and distribution of glycine in beef. On the basis of spectral information pre-processing, spectral features were extracted by the interval variable iterative space shrinkage approach, competitive adaptive reweighting algorithm, and uninformative variable elimination (UVE). The glycine content prediction models were established by partial least squares regression, least squares support vector machine, and the optimized shallow full convolutional neural network (SFCN). Among them, the UVE-SFCN model was found to show better results with prediction set determination coefficient (RP2) of 0.8725. Furthermore, textural features were extracted by the gray-level co-occurrence matrix and fused with the spectral information of the best feature band to obtain an optimized UVE-FSCN-fusion model (RP2 = 0.9005, root mean square error = 0.3075, residual predictive deviation = 0.2688). Compared with the full spectrum and characteristic wavelength spectrum models, RP2 was improved by 6.41% and 3.10%. The best fusion model was visualized to represent the distribution of glycine in beef. The results showed that the prediction and visualization of glycine content in beef were feasible and effective, and provided a theoretical basis for the hyperspectral study of meat quality monitoring or the establishment of an online platform.

Application of visible-near-infrared hyperspectral imaging technology coupled with wavelength selection algorithm for rapid determination of moisture content of soybean seeds

Moisture content is a crucial factor affecting the quality of soybean seeds. However, the determination of moisture content of soybean seeds is time-consuming and expensive. In this study, visible-near-infrared hyperspectral imaging technology (400–1000 nm) coupled with wavelength selection algorithm was applied to determine the moisture content of soybean seeds. Hyperspectral images of 96 soybean samples were obtained, and the sample set partitioning based on joint x-y distance algorithm was used to divide the calibration and prediction sets after removing outliers. Then, partial least squares regression (PLSR) models based on the original and preprocessing spectra were established, and the prediction effect of the original spectra was better than that of the preprocessing spectra. Five wavelength selection algorithms were used to select feature wavelengths to optimize the models further. Each wavelength selection algorithm was run 100 times independently to investigate its stability. The PLSR models were established based on the results of wavelength selection, and the prediction effects of all models were statistically analyzed. Results showed that the combination of interval variable iterative space shrinkage approach and successive projections algorithm (IVISSA-SPA) based on the original spectra was the most suitable model for the determination of moisture content of soybean seeds. The prediction accuracies of the IVISSA-SPA model were R2P = 0.9713 ± 0.0044, RMSEP = 0.307 ± 0.021 and RPD = 6.058 ± 0.344 in 100 independent experiments. Results indicated that visible-near-infrared hyperspectral imaging coupled with wavelength selection algorithm provided a rapid method for determining the moisture content of soybean seeds.

Nutrient content prediction and geographical origin identification of red raspberry fruits by combining hyperspectral imaging with chemometrics

The geographical origin and the important nutrient contents greatly affect the quality of red raspberry (RRB, Rubus idaeus L.), a popular fruit with various health benefits. In this study, a chemometrics-assisted hyperspectral imaging (HSI) method was developed for predicting the nutrient contents, including pectin polysaccharides (PPS), reducing sugars (RS), total flavonoids (TF) and total phenolics (TP), and identifying the geographical origin of RRB fruits. The results showed that these nutrient contents in RRB fruits had significant differences between regions (P < 0.05) and could be well predicted based on the HSI full or effective wavelengths selected through competitive adaptive reweighted sampling (CARS) and variable iterative space shrinkage approach (VISSA). The best prediction results of PPS, RS, TF, and TP contents were achieved with the highest residual predictive deviation (RPD) values of 3.66, 3.95, 2.85, and 4.85, respectively. The RRB fruits from multi-regions in China were effectively distinguished by using the first derivative-partial least squares discriminant analysis (DER-PLSDA) model, with an accuracy of above 97%. Meanwhile, the fruits from three protected geographical indication (PGI) regions were successfully classified by using the orthogonal partial least squares discrimination analysis (OPLSDA) model, with an accuracy of above 98%. The study results indicate that HSI assisted with chemometrics is a promising method for predicting the important nutrient contents and identifying the geographical origin of red raspberry fruits.

Nondestructive evaluation of Zn content in rape leaves using MSSAE and hyperspectral imaging

Zinc (Zn) content plays a decisive role in plant growth. Accurate management of Zn fertilizer application can promote high-quality development of the oilseed rape industry. This study adopted a deep learning (DL) method to predict the Zn content of oilseed rape leaves using hyperspectral imaging (HSI). The dropout mechanism was introduced to improve the stacked sparse autoencoder (SSAE) and named modified SSAE (MSSAE). MSSAE extracted deep spectral features of samples based on pixel-level spectral information (the wavelength range of the spectrum is 431–962 nm). Subsequently, the deep spectral features were applied as the inputs for support vector regression (SVR) and least squares support vector regression (LSSVR) to predict the Zn content in oilseed rape leaves. In addition, the successive projections algorithm (SPA) and the variable iterative space shrinkage approach (VISSA) were investigated as wavelength selection algorithms for comparison. The results showed that the MSSAE-LSSVR model had the best prediction performance (the coefficient of determination (R2) and root mean square error (RMSE) of the prediction set were 0.9566 and 1.0240 mg/kg, respectively). The overall results showed that the MSSAE was able to extract the deep features of HSI data and validated the possibility of HSI combined with a DL method for nondestructive testing of Zn content in oilseed rape leaves.

Rapid and Low-Cost Detection of Millet Quality by Miniature Near-Infrared Spectroscopy and Iteratively Retaining Informative Variables.

Traditional chemical methods for testing the fat content of millet, a widely consumed grain, are time-consuming and costly. In this study, we developed a low-cost and rapid method for fat detection and quantification in millet. A miniature NIR spectrometer connected to a smartphone was used to collect spectral data from millet samples of different origins. The standard normal variate (SNV) and first derivative (1D) methods were used to preprocess spectral signals. Variable selection methods, including bootstrapping soft shrinkage (BOSS), the variable iterative space shrinkage approach (VISSA), iteratively retaining informative variables (IRIV), iteratively variable subset optimization (IVSO), and competitive adaptive reweighted sampling (CARS), were used to select characteristic wavelengths. The partial least squares regression (PLSR) algorithm was employed to develop the regression models aimed at predicting the fat content in millet. The results showed that the proposed 1D-IRIV-PLSR model achieved optimal accuracy for fat detection, with a correlation coefficient for prediction (Rp) of 0.953, a root mean square error for prediction (RMSEP) of 0.301 g/100 g, and a residual predictive deviation (RPD) of 3.225, by using only 18 characteristic wavelengths. This result highlights the feasibility of using this low-cost and high-portability assessment tool for millet quality testing, which provides an optional solution for in situ inspection of millet quality in different scenarios, such as production lines or sales stores.

Rapid evaluation of texture parameters of Tan mutton using hyperspectral imaging with optimization algorithms

The detection of meat texture is of great value because it is the key factor that drives consumer purchasing decisions. In this study, a hyperspectral imaging (HSI) system was utilized to determine the texture parameters of Tan mutton. In order to observe the influence of mutton spectra during different refrigeration periods for modeling, hyperspectral images of the Tan mutton samples were collected in the 900–1700 nm spectral range, and the correction models of Tan mutton texture parameters were established. The four machine learning algorithms, such as partial least squares regression (PLSR), least squares support vector machine (LSSVM), random forest (RF), and decision trees (DT), were developed to establish the spectral models based on the characteristic bands selected by different extraction strategies including interval variable iterative space shrinkage approach (iVISSA), competitive adaptive reweighted sampling (CARS), successive projection algorithm (SPA) and variable combination population analysis (VCPA). The results showed that the LSSVM-iVISSA-CARS models exhibited excellent performance in predicting hardness and gumminess with root-mean-square errors (RMSEP) of 5.259 and 3.051 as well as the coefficient of determination for the prediction data set (R p 2 ) of 0.986 and 0.984 respectively. Good performances were achieved with R p 2 of 0.987 and RMSEP of 4.970 with the LSSVM-iVISSA-SPA model for chewiness, respectively. Therefore, HSI has potential for the evaluation and prediction of texture parameters in Tan mutton. • Spectra at 900–1700 nm was applied for detecting texture parameters of Tan mutton. • The OSC and MC methods were both utilized to preprocess X-block and Y-block. • Informative wavelengths were selected using iVISSV-CARS and iVISSV-SPA approaches. • The LSSVM model yielded the best result in predicting hardness, gumminess and chewiness of Tan mutton.

Characteristic wavelengths optimization improved the predictive performance of near-infrared spectroscopy models for determination of aflatoxin B1 in maize

A neoteric measure for quantitative assay of Aflatoxin B1 (AFB1) in maize based on an optimized feature model of near-infrared (NIR) spectroscopy was proposed in the work. A portable near-infrared spectroscopy system constructed by the group was employed to collect maize samples with varying degrees of mildew. The variable selection methods of interval variable iterative space shrinkage approach (IVISSA), iterative retained information variable (IRIV), and particle swarm optimization combined moving window (PSO-CMW) were introduced to perform feature selection on the pretreatment NIR spectra. The characteristic wavelength variables after screening were used to constitute support vector machine (SVM) and partial least squares (PLS) test model respectively to implement the measurement of AFB1 in maize, and the detection performance of the two types of models was compared. The results obtained showed that the overall performances of SVM models were higher than that of PLS models, and the SVM model based on the characteristic wavelength variables optimized by the PSO-CMW method had the most prominent generalization performance. The root mean square error of prediction (RMSEP) of the model was 3.5967 μg kg−1, the coefficient of determination (RP2) was 0.9707, and the relative prediction deviation (RPD) was 5.7538. The overall results demonstrate that the optimized features of NIR spectra can realize the on-site quick testing of the AFB1 in maize with high precision by constructing a nonlinear SVM detection model. This investigation provides an original approach for speedy quantitative detection of mycotoxins in cereals.

Non-Destructive Detection of pH Value of Kiwifruit Based on Hyperspectral Fluorescence Imaging Technology

Non-destructive detection of the pH value of kiwifruit has important practical significance for its quality classification. In this study, hyperspectral fluorescence imaging technology was proposed to quantitatively predict the pH value of kiwifruit non-destructively. Firstly, the SPXY algorithm was used to divide samples into training and prediction sets and three different algorithms were used to preprocess the raw spectral data. Secondly, algorithms such as the iteratively retaining information variables (IRIV), the variable iterative space shrinkage approach (VISSA), the model adaptive space shrinkage (MASS), the random frog (RF), and their combination (i.e., IRIV + VISSA + MASS + RF, IVMR) were used to extract effective variables from the preprocessed spectral data. Moreover, the second extractions, such as IRIV-VISSA and IRIV-MASS, and the third extraction (i.e., IVMR-VISSA-IRIV) were used to further reduce the redundant variables. Based on the effective variables, four regression models—random forest (RF), partial least square (PLSR), extreme learning machines (ELM), and multiple-kernel support vector regression (MK-SVR)—were built and compared for predicting. The results show that IVMR-VISSA-IRIV-MK-SVR had the best prediction results, with RP2, RC2 and RPD of 0.8512, 0.8580, and 2.66, respectively, which verifies that hyperspectral fluorescence imaging technology is reliable for predicting the pH value of kiwifruit non-destructively.

Detection of early bruises in jujubes based on reflectance, absorbance and Kubelka-Munk spectral data

Bruising is one of the major challenges appearing in the postharvest grading and processing of Lingwu long jujubes, which leads to quality deterioration and microbial infection. In the present work, the spectra of reflectance (R), absorbance (A) and Kubelka-Munk (K-M) in Lingwu were obtained by hyperspectral imaging, and applied to non-destructively detect bruising symptoms. Spectra were preprocessed and characteristic wavelengths were selected by competitive adaptive reweighted sampling (CARS) and the interval variable iterative space shrinkage approach (iVISSA). Classification discriminant models were constructed by partial least squares-discriminant analysis (PLS-DA) and support vector machine (SVM). By comparison, the results revealed that the A-raw-iVISSA-PLS-DA model showed the lowest errors in cross-validation, while the number of feature variables was the lowest accounting for 28.8 %, and the accuracies of the calibration and cross validation were 88.9 % and 100.0 %, respectively. In particular, this study demonstrated the feasibility to detect the early bruising degree in Lingwu long jujubes based on absorbance spectrum. Consequently, it also laid a foundation for future studies about detecting early bruising in small fruit with a rapid and non-destructive spectral-optical measurement.

Research on Classification Model of Panax notoginseng Taproots Based on Machine Vision Feature Fusion.

The existing classification methods for Panax notoginseng taproots suffer from low accuracy, low efficiency, and poor stability. In this study, a classification model based on image feature fusion is established for Panax notoginseng taproots. The images of Panax notoginseng taproots collected in the experiment are preprocessed by Gaussian filtering, binarization, and morphological methods. Then, a total of 40 features are extracted, including size and shape features, HSV and RGB color features, and texture features. Through BP neural network, extreme learning machine (ELM), and support vector machine (SVM) models, the importance of color, texture, and fusion features for the classification of the main roots of Panax notoginseng is verified. Among the three models, the SVM model performs the best, achieving an accuracy of 92.037% on the prediction set. Next, iterative retaining information variables (IRIVs), variable iterative space shrinkage approach (VISSA), and stepwise regression analysis (SRA) are used to reduce the dimension of all the features. Finally, a traditional machine learning SVM model based on feature selection and a deep learning model based on semantic segmentation are established. With the model size of only 125 kb and the training time of 3.4 s, the IRIV-SVM model achieves an accuracy of 95.370% on the test set, so IRIV-SVM is selected as the main root classification model for Panax notoginseng. After being optimized by the gray wolf optimizer, the IRIV-GWO-SVM model achieves the highest classification accuracy of 98.704% on the test set. The study results of this paper provide a basis for developing online classification methods of Panax notoginseng with different grades in actual production.

Rapid determination of TBARS content by hyperspectral imaging for evaluating lipid oxidation in mutton

This study aimed to investigate the feasibility of a rapid and non-destructive detecting method for lipid oxidation degree in mutton by near-infrared hyperspectral imaging (NIR-HSI). The hyperspectral images of the samples were collected in the spectral range of 900–1700 nm based on line-scanning. Partial least squares regression (PLSR) and least-squares support vector machines (LSSVM) were established to correlate the full spectra with measured TBARS values. To reduce the complexity of the calibration models, eight new PLSR and LSSVM models based on the four sets of key variables extracted by interval variable iterative space shrinkage approach (iVISSA), interval random frog (IRF), variable combination population analysis (VCPA) and competitive adaptive reweighted sampling (CARS) were developed and compared. The optimized LSSVM model based on the feature wavebands selected by CARS methods was found to be good in predicting TBARS content, with Rp2 of 0.83, RMSEP of 0.11 mg kg−1, RPD of 2.82 and RER of 10.36. In summary, this study showed that it was feasible to non-destructively and rapidly evaluate lipid oxidation degree in mutton using NIR-HSI.

Comparison of wavelength selected methods for improving of prediction performance of PLS model to determine aflatoxin B1 (AFB1) in wheat samples during storage

Wheat is a widely grown grain crop around the world and is highly susceptible to environmental factors during storage and transportation, resulting in the production of fungal toxins that are harmful to humans. Of these, aflatoxin B1 (AFB1) is the most prevalent and most toxic. In view of this, this study used a self-built portable near-infrared spectroscopy system to predict the AFB1 content of wheat during storage and investigated and compared the prediction effects of different wavelength selection algorithms on the constructed PLS model. Firstly, the NIR spectra of wheat samples at disparate storage stages were acquired using the NIR spectroscopy system. Secondly, the raw NIR spectra were pretreated by Savizkg-Golag (SG) smoothing, standard normal variate (SNV) and normalization in turn. Finally, three variable optimization methods, which were variable combination population analysis (VCPA), variable iterative space shrinkage approach (VISSA) and competitive adaptive reweighted sampling (CARS), were applied to select the characteristic wavelength variables of the pre-processed spectra. Partial least squares (PLS) models based on the optimized features of the three methods were established, respectively. The results obtained showed that the CARS-PLS model had the best overall effect. The root mean square error of prediction (RMSEP) for the best CARS-PLS was 2.0965 μg∙kg-1, the prediction coefficient of determination (Rp2) was 0.9935, and the ratio of prediction to deviation (RPD) was 7.3279. The CARS variable screening method was used to effectively select the characteristic wavebands associated with AFB1 in wheat, compressing the number of wavelength variables, simplifying the model structure and improving model performance. The results reveal that the self-built portable NIR spectroscopy system enables to determine the AFB1 in wheat during storage. Furthermore, through the feature optimization of spectral wavelength variables can effectively exclude undesired wavelength variable information.