Non-destructive detection and classification of textile fibres based on hyperspectral imaging and 1D-CNN

Non-destructive prediction of TVB-N using color-texture features of UV-induced fluorescence image for freeze-thaw treated frozen-whole-round tilapia

BACKGROUND

The investigation of UV-induced fluorescence imaging coupled with machine learning was conducted to non-destructively detect the total volatile basic nitrogen (TVB-N) of frozen-whole-round tilapia (FWRT) during freezing and thawing. The UV-induced fluorescence images of FWRT at the wavelength of 365 nm were acquired by self-developed fluorescence image acquisition system. In total, 169 color and texture features based on RGB, hue-saturation-intensity and L*a*b* color spaces and gray level co-occurrence matrix were extracted, respectively. Successive projections algorithm (SPA) was employed to select the optimal 16 features to achieve feature dimension reduction modeling. With full and extracted features as input, the models of partial least squares regression (PLSR), least-squares support vector machine (LSSVM) and convolutional neural network (CNN) were established for TVB-N prediction.

RESULTS

Results indicated that the full features-based CNN performed better than SPA based prediction models (SPA-PLSR and SPA-LSSVM). The CNN model was determined to be the optimal with an RP2 value of 0.9779, RMSEP value of 1.1502 × 10-2  g N kg-1 and RPD value of 6.721 for TVB-N content predictiin.

CONCLUSION

The CNN method based on UV fluorescence imaging technology has potential for quality and safety detection of FWRT. © 2023 Society of Chemical Industry.

Establishment and comparison of in situ detection models for foodborne pathogen contamination on mutton based on SWIR-HSI

Introduction

Rapid and accurate detection of food-borne pathogens on mutton is of great significance to ensure the safety of mutton and its products and the health of consumers.

Objectives

The feasibility of short-wave infrared hyperspectral imaging (SWIR-HSI) in detecting the contamination status and species of Escherichia coli (EC), Staphylococcus aureus (SA) and Salmonella typhimurium (ST) contaminated on mutton was explored.

Materials and methods

The hyperspectral images of uncontaminated and contaminated mutton samples with different concentrations (108, 107, 106, 105, 104, 103 and 102 CFU/mL) of EC, SA and ST were acquired. The one dimensional convolutional neural network (1D-CNN) model was constructed and the influence of structure hyperparameters on the model was explored. The effects of different spectral preprocessing methods on partial least squares-discriminant analysis (PLS-DA), support vector machine (SVM) and 1D-CNN models were discussed. In addition, the feasibility of using the characteristic wavelength to establish simplified models was explored.

Results and discussion

The best full band model was the 1D-CNN model with the convolution kernels number of (64, 16) and the activation function of tanh established by the original spectra, and its accuracy of training set, test set and external validation set were 100.00, 92.86 and 97.62%, respectively. The optimal simplified model was genetic algorithm optimization support vector machine (GA-SVM). For discriminating the pathogen species, the accuracies of SVM models established by full band spectra preprocessed by 2D and all 1D-CNN models with the convolution kernel number of (32, 16) and the activation function of tanh were 100.00%. In addition, the accuracies of all simplified models were 100.00% except for the 1D-CNN models. Considering the complexity of features and model calculation, the 1D-CNN models established by original spectra were the optimal models for pathogenic bacteria contamination status and species. The simplified models provide basis for developing multispectral detection instruments.

Conclusion

The results proved that SWIR-HSI combined with machine learning and deep learning could accurately detect the foodborne pathogen contamination on mutton, and the performance of deep learning models were better than that of machine learning. This study can promote the application of HSI technology in the detection of foodborne pathogens on meat.

Classification and rapid non-destructive quality evaluation of different processed products of Cyperus rotundus based on near-infrared spectroscopy combined with deep learning

The quality of traditional Chinese medicine is very important for human health, but the traditional quality control method is very tedious, which leads to the substandard quality of many traditional Chinese medicine. In order to solve the problem of time-consuming and laborious traditional quality control methods, this study takes traditional Chinese medicine Cyperus rotundus as an example, a comprehensive strategy of near-infrared (NIR) spectroscopy combined with One-dimensional convolutional neural network (1D-CNN) and chaotic map dung beetle optimization (CDBO) algorithm combined with BP neural network (BPNN) is proposed. This strategy has the advantages of fast and non-destructive. It can not only qualitatively distinguish Cyperus rotundus and various processed products, but also quantitatively predict two bioactive components. In classification, 1D-CNN successfully distinguished four kinds of processed products of Cyperus rotundus with 100 % accuracy. Quantitatively, a CDBO algorithm is proposed to optimize the performance of the BPNN quantitative model of two terpenoids, and compared with the BP, whale optimization algorithm (WOA)-BP, sparrow optimization algorithm (SSA)-BP, grey wolf optimization (GWO)-BP and particle swarm optimization (PSO)-BP models. The results show that the CDBO-BPNN model has the smallest error and has a significant advantage in predicting the content of active components in different processed products. To sum up, it is feasible to use near infrared spectroscopy to quickly evaluate the effect of processing methods on the quality of Cyperus rotundus, which provides a meaningful reference for the quality control of traditional Chinese medicine with many other processing methods.

Prediction of soluble solids content using near-infrared spectra and optical properties of intact apple and pulp applying PLSR and CNN

Soluble solids content (SSC) is one of the most important internal quality attributes of fruit and could be predicted using near-infrared (NIR) spectra and optical properties. Partial least squares regression (PLSR) is a conventional regression method in SSC prediction. In recent years, deep learning methods represented by convolutional neural network (CNN) was suggested to be implied in spectral analysis. However, researchers are inevitably facing problems with regard to the selection of spectral pretreatment methods and the evaluation of the performance of the chosen regression. This study employed PLSR and CNN regression to predict SSC of apple based on the collected diffuse reflectance spectra of intact apple, total reflectance and total transmittance spectra of apple pulp, and the calculated optical property spectra, i.e., absorption coefficient and reduced scattering coefficient spectra of apple pulp. Five different spectral pretreatment methods were exerted on these spectra. Results showed that at a given regression (PLSR or CNN), the built models based on the diffuse reflectance spectra of intact apple had the best SSC prediction, and the built models based on pulp's reduced scattering coefficient spectra had the poorest prediction performance. The best prediction performance was achieved by PLSR models using Savitzky-Golay with multiple scattering correction (Rp = 0.96, RMSEP = 0.54 %) and CNN regressions using Savitzky-Golay with standard normal variational transformation (Rp = 0.95, RMSEP = 0.59 %), respectively. Additionally, when the unknown original spectra were used for modeling, CNN had a better performance compared to PLSR, indicating the outstanding preponderance of CNN in spectral analysis. This study provides an effective reference for the selection of chemometric method based on NIR spectra.

A Nondestructive Methodology for Determining Chemical Composition of Salvia miltiorrhiza via Hyperspectral Imaging Analysis and Squeeze-and-Excitation Residual Networks

The quality assurance of bulk medicinal materials, crucial for botanical drug production, necessitates advanced analytical methods. Conventional techniques, including high-performance liquid chromatography, require extensive pre-processing and rely on extensive solvent use, presenting both environmental and safety concerns. Accordingly, a non-destructive, expedited approach for assessing both the chemical and physical attributes of these materials is imperative for streamlined manufacturing. We introduce an innovative method, designated as Squeeze-and-Excitation Residual Network Combined Hyperspectral Image Analysis (SE-ReHIA), for the swift and non-invasive assessment of the chemical makeup of bulk medicinal substances. In a demonstrative application, hyperspectral imaging in the 389-1020 nm range was employed in 187 batches of Salvia miltiorrhiza. Notable constituents such as salvianolic acid B, dihydrotanshinone I, cryptotanshinone, tanshinone IIA, and moisture were quantified. The SE-ReHIA model, incorporating convolutional layers, maxpooling layers, squeeze-and-excitation residual blocks, and fully connected layers, exhibited Rc2 values of 0.981, 0.980, 0.975, 0.972, and 0.970 for the aforementioned compounds and moisture. Furthermore, Rp2 values were ascertained to be 0.975, 0.943, 0.962, 0.957, and 0.930, respectively, signifying the model's commendable predictive competence. This study marks the inaugural application of SE-ReHIA for Salvia miltiorrhiza's chemical profiling, offering a method that is rapid, eco-friendly, and non-invasive. Such advancements can fortify consistency across botanical drug batches, underpinning product reliability. The broader applicability of the SE-ReHIA technique in the quality assurance of bulk medicinal entities is anticipated with optimism.

Comparison of Performance of Predicting the Wear Amount of Tire Tread Depending on Sensing Information

Excessive tire wear can affect vehicle driving safety. While there are various methods for predicting the tire wear amount in real-time, it is unclear which method is the most effective in terms of the difficulty of sensing and prediction accuracy. The current study aims to develop prediction algorithms of tire wear and compare their performances. A finite element tire model was developed and validated against experimental data. Parametric tire rolling simulations were conducted using various driving and tire wear conditions to obtain tire internal accelerations. Machine-learning-based algorithms for tire wear prediction utilizing various sensing options were developed, and their performances were compared. A wheel translational and rotational speed-based (V and ω) method resulted in an average prediction error of 1.2 mm. Utilizing the internal pressure and vertical load of the tire with the V and ω improved the prediction accuracy to 0.34 mm. Acceleration-based methods resulted in an average prediction error of 0.6 mm. An algorithm using both the vehicle and tire information showed the best performance with a prediction error of 0.21 mm. When accounting for sensing cost, the V and ω-based method seems to be promising option. This finding needs to be experimentally verified.