Detection of Aspergillus spp. contamination levels in peanuts by near infrared spectroscopy and electronic nose

Non-destructive detection of defective maize kernels using hyperspectral imaging and convolutional neural network with attention module

Rapid, effective and non-destructive detection of the defective maize kernels is crucial for their high-quality storage in granary. Hyperspectral imaging (HSI) coupled with convolutional neural network (CNN) based on spectral and spatial attention (Spl-Spal-At) module was proposed for identifying the different types of maize kernels. The HSI data within 380-1000 nm of six classes of sprouted, heat-damaged, insect-damaged, moldy, broken and healthy kernels was collected. The CNN-Spl-At, CNN-Spal-At and CNN-Spl-Spal-At models were established based on the spectra, images and their fusion features as inputs for the recognition of different kernels. Further compared the performances of proposed models and conventional models were built by support vector machine (SVM) and extreme learning machine (ELM). The results indicated that the recognition ability of CNN with attention series models was significantly better than that of SVM and ELM models and fused features were more conducive to expressing the appearance of different kernels than single features. And the CNN-Spl-Spal-At model had an optimal recognition result with high average classification accuracy of 98.04 % and 94.56 % for the training and testing sets, respectively. The recognition results were visually presented on the surface image of kernels with different colors. The CNN-Spl-Spal-At model was built in this study could effectively detect defective maize kernels, and it also had great potential to provide the analysis approaches for the development of non-destructive testing equipment based on HSI technique for maize quality.

Fusing hyperspectral imaging and electronic nose data to predict moisture content in Penaeus vannamei during solar drying

The control of moisture content (MC) is essential in the drying of shrimp, directly impacting its quality and shelf life. This study aimed to develop an accurate method for determining shrimp MC by integrating hyperspectral imaging (HSI) with electronic nose (E-nose) technology. We employed three different data fusion approaches: pixel-, feature-, and decision-fusion, to combine HSI and E nose data for the prediction of shrimp MC. We developed partial least squares regression (PLSR) models for each method and compared their performance in terms of prediction accuracy. The decision fusion approach outperformed the other methods, producing the highest determination coefficients for both calibration (0.9595) and validation sets (0.9448). Corresponding root-mean square errors were the lowest for the calibration set (0.0370) and validation set (0.0443), indicating high prediction precision. Additionally, this approach achieved a relative percent deviation of 3.94, the highest among the methods tested. The findings suggest that the decision fusion of HSI and E nose data through a PLSR model is an effective, accurate, and efficient method for evaluating shrimp MC. The demonstrated capability of this approach makes it a valuable tool for quality control and market monitoring of dried shrimp products.

Recent developments of e-sensing devices coupled to data processing techniques in food quality evaluation: a critical review

A greater demand for high-quality food is being driven by the growth of economic and technological advancements. In this context, consumers are currently paying special attention to organoleptic characteristics such as smell, taste, and appearance. Motivated to mimic human senses, scientists developed electronic devices such as e-noses, e-tongues, and e-eyes, to spot signals relative to different chemical substances prevalent in food systems. To interpret the information provided by the sensors' responses, multiple chemometric approaches are used depending on the aim of the study. This review based on the Web of Science database, endeavored to scrutinize three e-sensing systems coupled to chemometric approaches for food quality evaluation. A total of 122 eligible articles pertaining to the e-nose, e-tongue and e-eye devices were selected to conduct this review. Most of the performed studies used exploratory analysis based on linear factorial methods, while classification and regression techniques came in the second position. Although their applications have been less common in food science, it is to be noted that nonlinear approaches based on artificial intelligence and machine learning deployed in a big-data context have generally yielded better results for classification and regression purposes, providing new perspectives for future studies.

A Vis/NIR spectra-based approach for identifying bananas infected with Colletotrichum musae

Introduction

Anthracnose of banana caused by Colletotrichum species is one of the most serious post-harvest diseases, which can cause significant yield losses. Clarifying the infection mechanism of the fungi using non-destructive methods is crucial for timely discriminating infected bananas and taking preventive and control measures.

Methods

This study presented an approach for tracking growth and identifying different infection stages of the C. musae in bananas using Vis/NIR spectroscopy. A total of 330 banana reflectance spectra were collected over ten consecutive days after inoculation, with a sampling rate of 24 h. The four-class and five-class discriminant patterns were designed to examine the capability of NIR spectra in discriminating bananas infected at different levels (control, acceptable, moldy, and highly moldy), and different time at early stage (control and days 1-4). Three traditional feature extraction methods, i.e. PC loading coefficient (PCA), competitive adaptive reweighted sampling (CARS) and successive projections algorithm (SPA), combining with two machine learning methods, i.e. partial least squares discriminant analysis (PLSDA) and support vector machine (SVM), were employed to build discriminant models. One-dimensional convolutional neural network (1D-CNN) without manually extracted feature parameters was also introduced for comparison.

Results

The PCA-SVM and·SPA-SVM models had good performance with identification accuracies of 93.98% and 91.57%, 94.47% and 89.47% in validation sets for the four- and five-class patterns, respectively. While the 1D-CNN models performed the best, achieving an accuracy of 95.18% and 97.37% for identifying infected bananas at different levels and time, respectively.

Discussion

These results indicate the feasibility of identifying banana fruit infected with C. musae using Vis/NIR spectra, and the resolution can be accurate to one day.

Spatio-temporal distribution patterns and quantitative detection of aflatoxin B1 and total aflatoxin in peanut kernels explored by short-wave infrared hyperspectral imaging

Aflatoxin contamination in peanut kernels seriously harms the health of humans and causes significant economic losses. Rapid and accurate detection of aflatoxin is necessary to minimize its contamination. However, current detection methods are time-consuming, expensive and destructive to samples. Therefore, short-wave infrared (SWIR) hyperspectral imaging coupled with multivariate statistical analysis was used to investigate the spatio-temporal distribution patterns of aflatoxin, and quantitatively detect the aflatoxin B₁ (AFB₁) and total aflatoxin in peanut kernels. In addition, Aspergillus flavus contamination was identified to prevent the production of aflatoxin. The result of validation set demonstrated that SWIR hyperspectral imaging could predict the contents of the AFB₁ and total aflatoxin accurately, with residual prediction deviation values of 2.7959 and 2.7274, and limits of detection of 29.3722 and 45.7429 μg/kg, respectively. This study presents a novel method for the quantitative detection of aflatoxin and offers an early warning system for its potential application.

Accurate and non-destructive monitoring of mold contamination in foodstuffs based on whole-cell biosensor array coupling with machine-learning prediction models

Mold contamination in foodstuffs causes huge economic losses, quality deterioration and mycotoxin production. Thus, non-destructive and accurate monitoring of mold occurrence in foodstuffs is highly required. We proposed a novel whole-cell biosensor array to monitor pre-mold events in foodstuffs. Firstly, 3 volatile markers ethyl propionate, 1-methyl-1 H-pyrrole and 2,3-butanediol were identified from pre-mold peanuts using gas chromatography-mass spectrometry. Together with other 3 frequently-reported volatiles from Aspergillus flavus infection, the volatiles at subinhibitory concentrations induced significant but differential response patterns from 14 stress-responsive Escherichia coli promoters. Subsequently, a whole-cell biosensor array based on the 14 promoters was constructed after whole-cell immobilization in calcium alginate. To discriminate the response patterns of the whole-cell biosensor array to mold-contaminated foodstuffs, optimal classifiers were determined by comparing 6 machine-learning algorithms. 100 % accuracy was achieved to discriminate healthy from moldy peanuts and maize, and 95 % and 98 % accuracy in discriminating pre-mold stages for infected peanuts and maize, based on random forest classifiers. 83 % accuracy was obtained to separate moldy peanuts from moldy maize by sparse partial least square determination analysis. The results demonstrated high accuracy and practicality of our method based on a whole-cell biosensor array coupling with machine-learning classifiers for mold monitoring in foodstuffs.

Quantitative study of impact damage on yellow peaches based on reflectance, absorbance and Kubelka-Munk spectral data

Impact damage is one of the main forms of damage during the postharvest transportation and processing of yellow peaches. Thus, a quantitative prediction of the impact damage degree of yellow peaches is significant for their postharvest grading. In the present study, mechanical parameters such as the damage area, absorbed energy and maximum force were obtained based on a single pendulum collision device and an intelligent data acquisition system. The reflection spectra (R) of the damaged areas of yellow peaches were collected by a hyperspectral imaging system and transformed into absorbance (A) spectra and Kubelka-Munk (K-M) spectra. The R, A and K-M spectra were preprocessed by standard normal variables (SNV), moving average (MA) and Gaussian filtering (GF). Partial least squares regression (PLSR) models and support vector regression (SVR) models based on original and preprocessed spectra were established, respectively. By comparative analysis, the spectral data with better prediction performance (raw or preprocessed spectra) were selected from all spectra, and the characteristic wavelengths were selected by competitive adaptive reweighted sampling (CARS) and uninformative variable elimination (UVE). The PLSR and SVR models based on characteristic wavelengths were established. The results revealed that the prediction performance of the K-M-GF-CARS-PLSR model is the best. For the damage area, absorbed energy and maximum force, the R _P ² and RMSEP of the K-M-GF-CARS-PLSR model were 0.870 and 77.865 mm², 0.772 and 1.065 J, 0.895 and 47.996 N, respectively. Furthermore, the values of their RPD were 2.700, 1.768 and 3.050, respectively. The characteristic wavelengths of the model were 18.8%, 10.2% and 21.6%, respectively. The results of this study showed that there was a strong correlation between the mechanical parameters and K-M spectrum, which demonstrates the feasibility of quantitatively predicting the damage degree of yellow peaches based on the K-M spectrum. Therefore, the results of this work not only provide theoretical guidance for the postharvest grading of fruits, but also enrich the theoretical system of biomechanics.

Detecting Aflatoxin B1 in Peanuts by Fourier Transform Near-Infrared Transmission and Diffuse Reflection Spectroscopy

Aflatioxin B₁ (AFB₁) has been recognized by the International Agency of Research on Cancer as a group 1 carcinogen in animals and humans. A fast, batch, and real-time control and no chemical pollution method was developed for the discrimination and quantification prediction of AFB₁-infected peanuts by applying Fourier transform near-infrared (FT-NIR) coupled with chemometrics. Initially, the near-infrared transmission (NIRT) and diffuse reflection (NIRR) modules were applied to collect spectra of the samples. The principal component analysis (PCA) method was employed to extract the characteristic wavelength, followed by different preprocessing methods (seven methods) to build an effective linear discriminant analysis (LDA) classification and partial least squares (PLS) quantification models. The results showed that, for both the NIRT or NIRR modules, the LDA classification models satisfactorily distinguished peanuts infected with AFB₁ or from those not infected, with external validation showing a 100% correct identification rate and a 0% misjudgment rate. In addition, combined with the concentration of AFB₁ in peanuts determined by enzyme-linked immunoassay assay, the best partial least squares (PLS) models were established, with a combination of the first derivative and the Norris derivative filter smoothing pretreatment (R_c² = 0.937 and 0.984, RMSECV = 3.92% and 2.22%, RPD = 3.98 and 7.91 for NIRR and NIRT, respectively). The correlation coefficient between the predicted value and the reference value in the external verification was 0.998 and 0.917, respectively. This study highlights that both spectral acquisition modules meet the requirements of online, rapid, and accurate identification of peanut AFB₁ infection in the early stages.

Untargeted rapid differentiation and targeted growth tracking of fungal contamination in rice grains based on headspace-gas chromatography-ion mobility spectrometry

BACKGROUND

Milled rice are prone to be contaminated with spoilage or toxigenic fungi during storage, which may pose a real threat to human health. Most traditional methods require long periods of time for enumeration and quantification. However, headspace-gas chromatography-ion mobility spectrometry (HS-GC-IMS) technology could characterize the complex volatile organic compounds (VOCs) released from samples in a non-destructive and environmentally friendly manner. Thus, this study described an innovative HS-GC-IMS strategy for analyzing VOC profiles to detect fungal contamination in milled rice.

RESULTS

A total of 24 typical target compounds were identified. Analysis of variance-partial least squares regression (APLSR) showed significant correlations between the target compounds and colony counts of fungi. While the changes of selected volatile components (acetic acid, 3-hydroxy-2-butanone and oct-en-3-ol) in fungi-inoculated rice had sufficiently high positive correlations with the colony counts, the logistic model could effectively be used to monitor the growth of individual fungus (R²  = 0.902-0.980). PLSR could effectively be used to predict fungal colony counts in rice samples (R²  = 0.831-0.953), and the different fungi-inoculated rice samples at 24 h could be successfully distinguished by support vector machine (SVM) (94.6%). The ability of HS-GC-IMS to monitor fungal infection would help to prevent contaminated rice grains from entering the food chain.

CONCLUSIONS

This result indicated that HS-GC-IMS three-dimensional fingerprints may be appropriate for the early detection of fungal infection in rice grains. © 2021 Society of Chemical Industry.

A Preliminary Study to Classify Corn Silage for High or Low Mycotoxin Contamination by Using near Infrared Spectroscopy

Mycotoxins should be monitored in order to properly evaluate corn silage safety quality. In the present study, corn silage samples (n = 115) were collected in a survey, characterized for concentrations of mycotoxins, and scanned by a NIR spectrometer. Random Forest classification models for NIR calibration were developed by applying different cut-offs to classify samples for concentration (i.e., μg/kg dry matter) or count (i.e., n) of (i) total detectable mycotoxins; (ii) regulated and emerging Fusarium toxins; (iii) emerging Fusarium toxins; (iv) Fumonisins and their metabolites; and (v) Penicillium toxins. An over- and under-sampling re-balancing technique was applied and performed 100 times. The best predictive model for total sum and count (i.e., accuracy mean ± standard deviation) was obtained by applying cut-offs of 10,000 µg/kg DM (i.e., 96.0 ± 2.7%) or 34 (i.e., 97.1 ± 1.8%), respectively. Regulated and emerging Fusarium mycotoxins achieved accuracies slightly less than 90%. For the Penicillium mycotoxin contamination category, an accuracy of 95.1 ± 2.8% was obtained by using a cut-off limit of 350 µg/kg DM as a total sum or 98.6 ± 1.3% for a cut-off limit of five as mycotoxin count. In conclusion, this work was a preliminary study to discriminate corn silage for high or low mycotoxin contamination by using NIR spectroscopy.

Monitoring the growth of Fusarium graminearum in wheat kernels using multispectral imaging with chemometric methods

Wheat is an important agricultural economic crop providing energy and nutrition for human beings. However, wheat kernels are easily contaminated with Fusarium graminearum that is harmful to human health. In this study, a rapid and nondestructive detection method has been developed to identify the degree of contamination and determine the count of Fusarium graminearum in wheat kernels using multispectral imaging technology. Based on genetic algorithm (GA) and principal component analysis (PCA) data preprocessing methods combined with partial least squares (PLS), support vector machine (SVM) and back propagation neural network (BPNN) chemometric methods, identification and quantitative determination models were established. The best result was obtained by GA-BPNN with an accuracy of up to 100% in the identification of the degree of contamination in wheat kernels at different contamination periods. Comparison of the results from different methods revealed that the best prediction of the count of Fusarium graminearum was obtained by GA-SVM with the correlation coefficient (R) in the calibration set and prediction set being 0.9663 and 0.9292, while the root mean square error (RMSE) in the calibration set and prediction set was 0.5992 and 0.6725 CFU g^-1, respectively. It can be concluded that the combination of multispectral imaging and chemometric methods was potentially useful for rapid and nondestructive detection of cereal fungi in practice.

Early Detection of Mold-Contaminated Peanuts Using Machine Learning and Deep Features Based on Optical Coherence Tomography

Fungal infection is a pre-harvest and post-harvest crisis for farmers of peanuts. In environments with temperatures around 28 °C to 30 °C or relative humidity of approximately 90%, mold-contaminated peanuts have a considerable likelihood to be infected with Aflatoxins. Aflatoxins are known to be highly carcinogenic, posing danger to humans and livestock. In this work, we proposed a new approach for detection of mold-contaminated peanuts at an early stage. The approach employs the optical coherence tomography (OCT) imaging technique and an error-correcting output code (ECOC) based Support Vector Machine (SVM) trained on features extracted using a pre-trained Deep Convolutional Neural Network (DCNN). To this end, mold-contaminated and uncontaminated peanuts were scanned to create a data set of OCT images used for training and evaluation of the ECOC-SVM model. Results showed that the proposed approach is capable of detecting mold-contaminated peanuts with respective accuracies of approximately 85% and 96% after incubation periods of 48 and 96 h.

Targeted versus Nontargeted Green Strategies Based on Headspace-Gas Chromatography-Ion Mobility Spectrometry Combined with Chemometrics for Rapid Detection of Fungal Contamination on Wheat Kernels

Conventional methods for detecting fungal contamination are generally time-consuming and sample-destructive, making them impossible for large-scale nondestructive detection and real-time analysis. Therefore, the potential of headspace-gas chromatography-ion mobility spectrometry (HS-GC-IMS) was examined for the rapid determination of fungal infection on wheat samples in a rapid and nondestructive manner. In addition, the validation experiment of detecting the percent A. flavus infection presented in simulated field samples was carried out. Because the dual separation of HS-GC-IMS could generate massive amounts of three-dimensional data, proper chemometric processing was required. In this study, two chemometric strategies including: (i) nontargeted spectral fingerprinting and (ii) targeted specific markers were introduced to evaluate the performances of classification and prediction models. Results showed that satisfying results for the differentiation of fungal species were obtained based on both strategies (>80%) by the genetic algorithm optimized support vector machine (GA-SVM), and better values were obtained based on the first strategy (100%). Likewise, the GA-SVM model based on the first strategy achieved the best prediction performances (R² = 0.979-0.998) of colony counts in fungal infected samples. The results of validation experiment showed that GA-SVM models based on the first strategy could still provide satisfactory classification (86.67%) and prediction (R² = 0.889) performances for percent A. flavus infection presented in simulated field samples at day 4. This study indicated the feasibility of HS-GC-IMS-based approaches for the early detection of fungal contamination in wheat kernels.

Pixel-level aflatoxin detecting in maize based on feature selection and hyperspectral imaging

Aflatoxin is highly toxic and is easily found in maize, a little aflatoxin can induce liver cancer. In this paper, we used hyperspectral data in the pixel-level to build the aflatoxin classifying model, each of the pixel have 600 hyperspectral bands and labeled 'clean' or 'contaminated'. We use 3 method to extracted feature bands, one method is to select 4 hyperspectral bands from other articles: 390 nm, 440 nm, 540 nm and 710 nm, another method is to use feature extraction PCA to obtain first 5 pcs to shrink the hyperspectral volume, the third method is to use Fscnca, Fscmrmr, Relieff and Fishier algorithm to select top 10 feature bands. After feature band selection or extraction, we put the feature bands into Random Forest (RF) and K-nearest neighbor (KNN) to classify whether a pixel is polluted by aflatoxin. The highest accurate for feature selection is Relieff, it reached the accuracy of 99.38% with RF classifier and 98.77% in KNN classifier. PCA feature extraction with RF classifier also reached a high accuracy 93.83%. And the 600 bands without feature extraction reached the accuracy of 100%. Feature bands selected from other papers could reach an accuracy of 89.51%. The result shows that the feature extraction performs well on its own data set. And if the computing time is not taken into account, we could use full band to classify the aflatoxin due to its high accuracy.

On-line prediction of hazardous fungal contamination in stored maize by integrating Vis/NIR spectroscopy and computer vision

This work presents a fusion scheme to combine visible/near infrared spectroscopy and computer vision for on-line detection of Aspergillus spp. and Fusarium spp. contamination in stored maize. Spectroscopy and image information of 270 groups of maize kernels were collected at speed of 0.15 m/s. Principal component analysis indicated fungi growth on maize could be monitored by both techniques. Spectroscopy method was found sensitive for infection level identification, while computer vision was more effective for fungal strain recognition. Linear discriminant analysis based on fusion of spectral and image features provided 100% accuracy for discrimination of samples infected by different strains after stored for 12 d, which is at least 5.6% higher than single-type features. Classification rate of samples with different infection levels achieved 92.2%, also 5.5% and 10.0% higher than single technique. Moreover, data fusion improved colony counts prediction in samples by partial least squares regression, with root mean-square error of prediction value being reduced by 25.0% and 17.4%, respectively. This study demonstrated the superiority of data fusion for fungal detection in grain during on-line processing.

Comprehensive comparison of multiple quantitative near-infrared spectroscopy models for Aspergillus flavus contamination detection in peanut

BACKGROUND

Aspergillus flavus is a major pollutant in moldy peanuts, and it has a large influence on the taste of food. The secondary metabolites of Aspergillus flavus, including aflatoxin B1 (AFB1) and aflatoxin B2 (AFB2), are highly toxic and can expose humans to high risk. The total mold count (TMC) is an important index to determine the contamination degree and hygiene quality of peanut.

RESULTS

Quantitative calibration models were established based on full-band wavelengths and characteristic wavelengths, combined with chemometric methods, to explore the feasibility of the use of near-infrared spectroscopy (NIRS) for rapid detection of the TMC in peanuts. The successive projection algorithm (SPA) and elimination of uninformative variables (UVE) algorithms were used to extract the characteristic wavelengths. In comparison, the model built by original spectrum, selected with the UVE algorithm, gave the best result, with a correlation coefficient in a prediction set (R_P ) of 0.9577, a root mean square error for the prediction set (RMSEP) of 0.2336 Log CFU/g, and a residual predictive deviation (RPD) of 3.5041.

CONCLUSIONS

The results showed that NIRS is a rapid, practicable method for the quantitative detection of peanut Aspergillus flavus contamination. It is a promising method for detecting moldy peanuts and increasing peanut safety. © 2019 Society of Chemical Industry.

Grading and Sorting of Grape Berries Using Visible-Near Infrared Spectroscopy on the Basis of Multiple Inner Quality Parameters

The potential of visible-near infrared (vis/NIR) spectroscopy (400 nm to 1100 nm) for classification of grape berries on the basis of multi inner quality parameters was investigated. Stored Vitis vinifera L. cv. Manicure Finger and Vitis vinifera L. cv. Ugni Blanc grape berries were separated into three classes based on the distribution of total soluble solid content (SSC) and total phenolic compounds (TP). Partial least squares regression (PLS) was applied to predict the quality parameters, including color space CIELAB, SSC, and TP. The prediction results showed that the vis/NIR spectrum correlated with the SSC and TP present in the intact grape berries with determination coefficient of prediction (R_P²) in the range of 0.735 to 0.823. Next, the vis/NIR spectrum was used to distinguish between berries with different SSC and TP concentrations using partial least squares discrimination analysis (PLS-DA) with >77% accuracy. This study provides a method to identify stored grape quality classes based on the spectroscopy and distributions of multiple inner quality parameters.

Early discrimination and growth tracking of Aspergillus spp. contamination in rice kernels using electronic nose

Early detection of Aspergillus spp. contamination in rice was investigated by electronic nose (E-nose) in this study. Sterilized rice artificially inoculated with three Aspergillus strains were subjected to GC-MS and E-nose analyses. Principle Component Analysis (PCA), Partial Least Squares Regression (PLSR), Back-propagation neural network (BPNN), Support Vector Machine (SVM) and Learning Vector Quantization (LVQ) were employed for qualitative classification and quantitative regression. GC-MS analysis revealed a significant correlation between the volatile compounds and total amounts/species of fungi. While X-axis barycenters of PC1 scores were significantly correlated with fungal counts, logistic model could be employed to simulate the growth of individual fungus (R² = 0.978-0.996). Fungal species and counts in rice could be classified and predicted by BPNN (96.4%) and PLSR (R² = 0.886-0.917), respectively. The results demonstrated that E-nose combined with BPNN might offer the feasibility for early detection of Aspergillus spp. contamination in rice.

Application of multiplexing fiber optic laser induced fluorescence spectroscopy for detection of aflatoxin B1 contaminated pistachio kernels

To explore the effect of signal acquisition way on screening ability, the multiplexing fiber optic laser induced fluorescence spectroscopy (LIFS) system with one-, two- and three-probe, were employed respectively to detect artificially aflatoxin B₁ (AFB₁, 5, 10, 20, 30, and 50 ppb) contaminated 300 pistachio kernels in this study. Compared to one- and two-probe modes, highest accuracy (≥97.0%) by support vector machine (SVM) employing 390-660 nm were obtained using three-probe, which also showed the most attractive precision (root mean square error of prediction (RMSEP) < 4.5 ppb) for AFB₁ by stepwise multiple linear regression (SMLR) using 174-1100 nm. These suggested that the effective collection of spatial information could improve the performance of model, and the three-probe LIFS had a preliminary feasibility for discriminating pistachios contaminated with low concentration of AFB₁. Further study on classifying naturally contaminated samples is needed to validate the applicability of this system.