Hyperspectral Imaging Technology Research Articles

Non-Destructive Assessment of Stone Heritage Weathering Types Based on Machine Learning Method Using Hyperspectral Data

Abstract. Stone cultural heritage is exposed to various environments, resulting in a diverse range of weathering types. The identification of these weathering types is vital for targeted conservation efforts. In this paper, a weathering type classification method based on hyperspectral imaging technology is proposed. Firstly, fresh sandstones are collected from Yungang Grottoes to carry out the simulated weathering experiments, including freeze-thaw cycles and wet-dry cycles with acid, alkali and salt solutions. Subsequently, the hyperspectral imaging system was used to collect the visible-near-infrared (VNIR) and short-wave infrared (SWIR) images of the sandstone samples with different weathering types and degrees. The surface spectral reflectance of sandstone samples with different weathering types were used as training data, with weathering types serving as the labels. Support vector machine (SVM), K-nearest neighbour (KNN), linear discriminant analysis (LDA) and random forest (RF) were used to establish weathering type classification models. The results show that the SVM model and LDA model based on both VNIR and SWIR spectra exhibit outstanding performance, with a best accuracy of 0.994. The framework proposed in this paper facilitates rapid and non-contact assessment of the weathering types of the superficial layers of stone cultural heritage, thereby supporting more targeted conservation work.

Rapid determination of starch and alcohol contents in fermented grains by hyperspectral imaging combined with data fusion techniques.

Starch and alcohol serve as pivotal indicators in assessing the quality of lees fermentation. In this paper, two hyperspectral imaging (HSI) techniques (visible-near-infrared (Vis-NIR) and NIR) were utilized to acquire separate HSI data, which were then fused and analyzed toforecast the starch and alcohol contents during the fermentation of lees. Five preprocessing methods were first used to preprocess the Vis-NIR, NIR, and the fused Vis-NIR and NIR data, after which partial least squares regression models were established to determine the best preprocessing method. Following, competitive adaptive reweighted sampling, successive projection algorithm, and principal component analysis algorithms were used to extract the characteristic wavelengths to accurately predict the starch and alcohol levels. Finally, support vector machine (SVM)-AdaBoost and XGBoost models were built based on the low-level fusion (LLF) and intermediate-level fusion (ILF) of single Vis-NIR and NIR as well as the fused data. The results showed that the SVM-AdaBoost model built using the LLF data afterpreprocessing by standard normalized variable was most accurate for predicting the starch content, with an of 0.9976 and a root mean square error of prediction (RMSEP) of 0.0992. The XGBoost model built using ILF data was most accurate for predicting the alcohol content, with an of 0.9969 and an RMSEP of 0.0605. In conclusion, the analysis of fused data from distinct HSI technologies facilitates rapid and precise determination of the starch and alcohol contents in fermented grains.

Automatic detection method of bladder tumor cells based on color and shape features

Bladder urothelial carcinoma is the most common malignant tumor disease in urinary system, and its incidence rate ranks ninth in the world. In recent years, the continuous development of hyperspectral imaging technology has provided a new tool for the auxiliary diagnosis of bladder cancer. In this study, based on microscopic hyperspectral data, an automatic detection algorithm of bladder tumor cells combining color features and shape features is proposed. Support vector machine (SVM) is used to build classification models and compare the classification performance of spectral feature, spectral and shape fusion feature, and the fusion feature proposed in this paper on the same classifier. The results show that the sensitivity, specificity, and accuracy of our classification algorithm based on shape and color fusion features are 0.952, 0.897, and 0.920, respectively, which are better than the classification algorithm only using spectral features. Therefore, this study can effectively extract the cell features of bladder urothelial carcinoma smear, thus achieving automatic, real-time, and noninvasive detection of bladder tumor cells, and then helping doctors improve the efficiency of pathological diagnosis of bladder urothelial cancer, and providing a reliable basis for doctors to choose treatment plans and judge the prognosis of the disease.

Rapid Detection of Endogenous Impurities in Walnuts Using Near-Infrared Hyperspectral Imaging Technology

Impurities in walnut-based products pose a significant risk to human health, necessitating the development of rapid and efficient methods for detecting endogenous impurities in walnut shell-breaking materials. To address this issue, this study aims to investigate the feasibility of near-infrared hyperspectral imaging (NIR-HSI) in conjunction with conventional chemometric methods and deep learning models. First, biochemical experiments demonstrated differences in the total phenolic and flavonoid contents of different walnut kernels. An improved support vector machine model based on the butterfly optimization algorithm was constructed for classification using an NIR dataset of endogenous impurities in walnuts, achieving an accuracy of 96.12%. In addition, this study proposed the WT-NIRSNet model, a deep neural network with an additional attention mechanism, and compared it to a traditional chemometric approach. The study demonstrated an accuracy of 99.03% on the test set using WT-NIRSNet to solve the fast classification problem of endogenous impurities in walnuts. It can achieve desirable classification results using only raw spectral data without complex preprocessing operations and is superior to traditional chemometric models. Therefore, the NIR-HSI technique combined with deep neural networks has a significant potential for detecting endogenous impurities in walnuts.

Quantitative prediction and visualization of matcha color physicochemical indicators using hyperspectral microscope imaging technology

Color physicochemical indicators of matcha are related to sensory quality, and the distributions of color physicochemical indicators can reflect the uniformity of matcha powder. This study aimed to explore the feasibility of using hyperspectral microscope imaging (HMI) technology to predict and visualize matcha color physicochemical indicators. The average spectra (400–998 nm) were extracted from the regions of interest (ROI). Subsequently, competitive adaptive reweighted sampling (CARS), interval random frog (iRF), successive projections algorithm (SPA), and the combination of iRF and SPA (iRF-SPA) were adopted to screen the characteristic spectra variables that were associated with specific color physicochemical indicators. Partial least squares (PLS) regression was applied to develop the calibration models for predicting the color physicochemical indicators. The results showed that the iRF-SPA-PLS models obtained the optimal prediction results, with predictive correlation coefficients (Rp) of 0.9262 for L*, 0.8826 for a*, 0.8583 for b*, 0.8243 for chlorophyll a, 0.7518 for chlorophyll b, and 0.8093 for chlorophyll total. Finally, the distribution maps of matcha color physicochemical indicators were visualized using these optimized models. The results suggested that the HMI technology coupled with chemometrics has enormous potential to predict and visualize the color physicochemical indicators of matcha, which could reflect the uniformity of matcha powder to ensure the stability of matcha quality. The approach presented in this study can also be applied to quality control of other powdery foods.

Hyperspectral imaging combined with spectral-imagery feature fusion convolutional neural network to discriminate different geographical origins of wolfberries

Geographical origin significantly influences the composition of Chinese wolfberries. Establishing a rapid and effective model for origin traceability is crucial. In this study, visible-near infrared (Vis-NIR) hyperspectral imaging technology combined with the proposed Spectral-Imagery Feature Fusion Convolutional Neural Network (S-IFCNN) was developed to identify the origin of wolfberries for the first time. The process began with segmenting adherent wolfberries using a combination of minimum transformation and the watershed algorithm, enabling the extraction of both spectral and image data. We compared various preprocessing methods, input data types, and convolutional kernel sizes to assess their impact on the performance of 1DCNN and 2DCNN models. The optimal configurations were then integrated into the S-IFCNN for a comprehensive evaluation. The study demonstrates the S-IFCNN model's capability in leveraging spectral and image data from hyperspectral imaging for accurately identifying the geographical origins of wolfberries, achieving an accuracy of 91.99%. This research not only highlights the potential applications of the S-IFCNN model but also provides a theoretical framework for the use of hyperspectral imaging in tracing the geographical origins of wolfberries.

Quality detection and variety classification of pecan seeds using hyperspectral imaging technology combined with machine learning

Pecan (Carya illinoinensis K.), a well-known dried seed and woody oil tree, faces challenges in its industry due to complex quality assessment methods and confusing varieties. These challenges have seriously hampered the development of a large-scale pecan deep processing industry. This work aimed to apply hyperspectral imaging technology (HSI) combined with machine learning to evaluate the quality of pecan seeds and perform variety classification. The samples of this work were composed of 19 varieties of pecan seeds, with 30 seeds per variety. After spectral preprocessing, spectral features were extracted from the spectral profiles using feature extraction methods. Back-propagation neural network models and partial least squares models were established to predict the contents of crude fat and moisture in pecan seeds. Predictions of the best models gave good results with R2-score of 0.887 for the crude fat model and 0.950 for the moisture model. Additionally, support vector machine models were developed to identify pecan varieties. The model achieved good results in 19 pecan varieties identification with accuracy of 0.965. In conclusion, the combination of HSI and machine learning could be an effective tool in improving the pecan industry and providing sustainable and efficient methods in the production of pecan seeds.

In Situ Nondestructive Detection of Nitrogen Content in Soybean Leaves Based on Hyperspectral Imaging Technology

In this paper, hyperspectral imaging technology, combined with chemometrics methods, was used to detect the nitrogen content of soybean leaves, and to achieve the rapid, non-destructive and in situ detection of the nitrogen content in soybean leaves. Soybean leaves under different fertilization treatments were used as the research object, and the hyperspectral imaging data and the corresponding nitrogen content data of soybean leaves at different growth stages were obtained. Seven spectral preprocessing methods, such as Savitzky–Golay smoothing (SG), first derivative (1-Der), and direct orthogonal signal correction (DOSC), were used to establish the quantitative prediction models for soybean leaf nitrogen content, and the quantitative prediction models of different spectral preprocessing methods for soybean leaf nitrogen content were analyzed and compared. On this basis, successive projections algorithm (SPA), genetic algorithm (GA) and random frog (RF) were employed to select the characteristic wavelengths and compress the spectral data. The results showed the following: (1) The full-spectrum prediction model of soybean leaf nitrogen content based on DOSC pretreatment was the best. (2) The PLS model of soybean leaf nitrogen content based on the five characteristic wavelengths had the best prediction performance. (3) The spatial distribution map of soybean leaf nitrogen content was generated in a pixel manner using the extracted five characteristic wavelengths and the DOSC-RF-PLS model. The nitrogen content level of soybean leaves can be quantified in a simple way; this provides a foundation for rapid in situ non-destructive detection and the spatial distribution difference detection of soybean leaf nitrogen. (4) The overall results illustrated that hyperspectral imaging technology was a powerful tool for the spatial prediction of the nitrogen content in soybean leaves, which provided a new method for the spatial distribution of the soybean nutrient status and the dynamic monitoring of the growth status.

Tracking moisture contents in the pollution layer on a composite insulator surface using hyperspectral imaging technology.

Electrical insulators used in transmission lines and outdoor substations are exposed to severe environmental pollution, which significantly increases the risk of power system failure, especially when the pollution layer is highly humid due to adverse weather conditions. The focus of this paper is to establish an effective method for assessing the moisture content (MC) in pollution layers as it serves as a crucial indicator for evaluating the risk of failure in insulators. Hyperspectral imaging (HSI) technology with a spectral range of 371.08-1037.89 nm was applied to determine significant changes in reflectance spectral characteristics in insulators during dynamic wetting and drying periods. Partial least squares regression (PLSR) models were utilized to evaluate the data presentation enhancement abilities of spectral transformation models and the data dimensionality reduction abilities of characteristic band selection methods. Furthermore, PLSR models were developed to calculate the MC along the pixel dimension to visually retrieve the dynamic wetting and drying processes of the pollution layer. The R-squared and root-mean-square error (RMSE) results in the cross-verification set and prediction set of the RE-RF(70%)-PLSR model with two characteristic bands with a wavelength of 543.28 nm and 848.01 nm were as follows: RCV2 = 0.9824, RMSECV = 0.0367, RP2 = 0.9818, RMSEP = 0.0369, respectively. This research contributes towards the visualization retrieval of the MC and offers an important technique for analyzing flashover evolution, optimizing insulator design, and preparing coating materials for insulators.

Predictive modeling of milk adulteration with urea content using the gray wolf optimization algorithm and long and short-term memory network model

Traditional methods for the determining of urea contaminants in milk, including chromatography, spectrophotometry, and electrochemical processes, have drawbacks such as long testing times and sample destruction. However, hyperspectral imaging technology provides a fast, easy-to-operate, and real-time alternative. In this study, five preprocessing methods are employed, including standard scalar, standard normal variational, first derivative, multivariate scattering correction algorithm, and Savitzky-Golay smoothing. To further enhance the accuracy of predicting urea content in milk, a prediction optimization method based on the gray wolf optimization algorithm for the long short-term memory network was proposed, using a competitive adaptive reweighted sampling algorithm feature wavelength selection algorithm. In the optimized model in the test set experiments, the decision coefficient was improved by 0.56% to 0.9906, and root mean square error was reduced by 7.69% to 0.4996 compared to the long short-term memory network model. This study not only provides a theoretical basis but also presents a fast and nondestructive detection method for accurately predicting urea content in milk.

In-field and non-destructive determination of comprehensive maturity index and maturity stages of Camellia oleifera fruits using a portable hyperspectral imager

To efficiently detect the maturity stages of Camellia oleifera fruits, this study proposed a non-invasive method based on hyperspectral imaging technology. First, a portable hyperspectral imager was used for the in-field image acquisition of Camellia oleifera fruits at three maturity stages, and ten quality indexes were measured as reference standards. Then, factor analysis was performed to obtain the comprehensive maturity index (CMI) by analyzing the change trends and correlations of different indexes. To reduce the high dimensionality of spectral data, the successive projection algorithm (SPA) was employed to select effective feature wavelengths. The prediction models for CMI, including partial least squares regression (PLSR), support vector regression (SVR), extreme learning machine (ELM), and convolutional neural network regression (CNNR), were constructed based on full spectra and feature wavelengths; for CNNR, only the raw spectra were used as input. The SPA-CNNR model exhibited more promising performance (RP = 0.839, RMSEP = 0.261, and RPD = 1.849). Furthermore, PLS-DA models for maturity discrimination of Camellia oleifera fruits were developed using full wavelength, characteristic wavelengths and their fusion CMI, respectively. The PLS-DA model using the fused dataset achieved the highest maturity classification accuracy, with the best simplified model achieving 88.6 % accuracy in prediction set. This study indicated that a portable hyperspectral imager can be used for in-field determination of the internal quality and maturity stages of Camellia oleifera fruits. It provides strong support for non-destructive quality inspection and timely harvesting of Camellia oleifera fruits in the field.

3D-GhostNet: A novel spatial-spectral algorithm to improve foodborne bacteria classification coupled with hyperspectral microscopic imaging technology

A novel 3D-GhostNet has been developed for hyperspectral microscopic imaging (HMI) data analysis to improve classification of foodborne pathogens. The HMI, leveraging microscopic technology, surpasses traditional spectral imaging in terms of sensitivity and resolution by utilizing visible/near-infrared spectroscopy integrated with high-resolution single-cell images. The newly constructed 3D-Ghost network directly processes and recognizes single-cell hypercube, enables HMI technology to become near-real-time detection by combining with automated data processing pipeline. 3D-GhostNet incorporates Ghost modules as its backbone and adds convolutional block attention module (CBAM) for adaptively extracting high-dimensional depth spatial-spectral information. In the task of identifying four different foodborne bacteria cells, our 3D-GhostNet achieved 100% accuracy, surpassing spectral-only-based (SOB) classifiers such as linear discriminant analysis (LDA, 89.7%), support vector machine (SVM, 93.9%), 1D convolutional neural network (1D-CNN, 98%), and long short-term memory (LSTM, 98.5%). Furthermore, our independent blind test results indicate that 3D-GhostNet remains robust for identifying complex mixtures of pathogens. As a label-free detection tool, 3D-GhostNet-assisted HMI technology is fully competent in classification of different foodborne pathogens, suggesting broad applications in food safety and quality.

Combining transfer learning and hyperspectral imaging to identify bruises of pears across different bruise types.

Mechanical bruise is one of the most crucial factors affecting the quality of pears, which has a huge influence on postharvest transportation, storage, and sale of pears. To rapidly detect early bruises of pears across different bruise types, hyperspectral imaging technology coupled with transfer learning methods was performed in this study. Two transfer learning methods, that is, transfer component analysis (TCA) and manifold embedded distribution alignment (MEDA), were applied for two tasks (impact bruise→crush bruise, crush bruise→impact bruise). Supporting vector machine (SVM) was set as a baseline to conduct analysis and comparison of the transferability of the models. The result showed that, for task 1 (impact bruise→crush bruise), MEDA and TCA-SVM model achieved a classification accuracy of 93.33% and 91.11% in target domain, individually. For task 2 (crush bruise→impact bruise), MEDA and TCA-SVM model achieved an accuracy of 88.89% and 85.19% in target domain, respectively. Both the two models improved the accuracy compared with SVM models (84.44% for task 1; 77.04% for task 2). Overall, the results indicated that transfer learning approaches could perform pear bruise detection across different bruise types. Hyperspectral imaging in combination with transfer learning methods is a promising possibility for the efficient and cost-saving field detection of fruit bruises among different bruise types. PRACTICAL APPLICATION: The production and export of pears are faced with problems of mechanical damage due to vibration, collision, impact, and other factors, which cause chemical changes in color, odor, and taste. Sometimes the bruise was too slight to be ignored which would infect with other fruits in the future. In this study, we used hyperspectral imaging combined with transfer learning method could detect these slight bruises caused by different factors. Distinguishing different types of damage can provide a reference for quick judgment of the process causing damage and take prompt measures to reduce economic losses.

Nondestructive detection of egg freshness based on a decision-level fusion method using hyperspectral imaging technology

Nondestructive detection of egg freshness based on a decision-level fusion method using hyperspectral imaging technology

Applications of hyperspectral imaging technology in the food industry

Applications of hyperspectral imaging technology in the food industry

Investigation of biofluorescence produced by the red king crab Paralithodes camtschaticus

AbstractBiofluorescence is widely documented in marine organisms, yet few studies exist for decapods. After observing the king crab Paralithodes camtschaticus biofluoresces, we carried out studies on two separately maintained groups of male crabs under controlled conditions. Hyperspectral imaging on Group 1 (n = 18) examined the exoskeleton, whereas fluorospectrometry examined the hemolymph of Group 2 animals (n = 19). Both groups were investigated for fluorescence signals before and after exposure to a live shipping transportation simulation. The spines, chelae, eyestalks and cervical grooves of the cardiac region of P. camtschaticus fluoresce in the green spectra (∼500 nm), while the arthrodial membranes of the joints fluoresce in the red spectrum (∼680 nm). After the shipping simulation, we observed a significant decrease in fluorescence in the eyestalks (p = 0.009), while the cervical grooves showed a less significant change (p = 0.01). The hemolymph examined with 21 excitation wavelengths (250–350 nm) emitted fluorescence in ∼400–550 nm spectrum. We found a significant increase (p < 0.05) in fluorescence for 16 excitation wavelengths after transport simulation. The results presented in this study indicate that king crab fluorescence changes due to external stimuli. Fluorospectroscopy or hyperspectral imaging technology may serve as an effective early indicator of preclinical stress in these commercially important decapods.

Development of a new hyperspectral imaging technology with autoencoder-assisted generative adversarial network for predicting the content of polyunsaturated fatty acids in red meat

The establishment of a comprehensive predictive model for red meat polyunsaturated fatty acids holds profound significance for the food industry. However, challenges, such as intricate features and low chemical content bestow complexity upon this endeavor. In the study, an autoencoder-assisted generative adversarial network (AE-GAN) was used to address the intricacies of generative models in regression operations. Following numerous iterations, the AE-GAN generated samples akin to the original data. Upon the incorporation of these generated samples into training, the test set R2 values of Support Vector Regression, Random Forest and Fully Convolutional Network witnessed respective enhancements of 0.1589, 0.1482 and 0.2998. The outcomes underscore the efficacy of this novel approach in ameliorating the challenges faced by generative models in regression tasks, thereby augmenting the model's generalizability.

Fast prediction of diverse rare ginsenoside contents in Panax ginseng through hyperspectral imaging assisted with the temporal convolutional network-attention mechanism (TCNA) deep learning

Combining hyperspectral imaging (HSI) with deep learning algorithms provides an effective and fast approach for evaluating the quality of food and agricultural by-products. This study comprehensively determined the quality of ginseng (Panax ginseng C. A. Meyer), an important medicinal and nutritional food, by evaluating the contents of diverse rare ginsenosides (RGs) using HSI technology. The results indicated that the combination of HSI with the deep learning temporal convolutional network-attention mechanism (TCNA) model achieved the best results in predicting the contents of six types of RGs (Rh1, Rh2, F1, Rg3, F4, and Rk1) simultaneously and effectively. Especially, the content detection of the six RGs based on the effective wavelengths showed that the TCNA model achieved coefficient of determination (R2) values above 0.890 and relative percentage deviation (RPD) values higher than 3.0, demonstrating excellent model performance. Meanwhile, the use of effective wavelengths makes the results of the TCNA model have better interpretability, and the simultaneous output of six RGs contents significantly improves prediction efficiency. The HSI assisted with the TCNA algorithm provides a rapid and effective detection approach for simultaneously predicting the content of diverse quality indicators. All these results will provide a new reference for developing convenient and rapid HSI equipment in the food and agricultural industry for direct and comprehensive quality inspection in markets in the future.

Establishment of deep learning model for the growth of tea cutting seedlings based on hyperspectral imaging technique

The shoot and root growth not only serve as important indicators for measuring the quality of tea cutting seedlings, but also can provide scientific basis for agricultural production and fine seedling management. Currently, manual weighing is the main method used to measure shoot and root growth, which is both destructive and inefficient. The study proposes a method for monitoring the shoot and root growth in tea cutting seedlings using hyperspectral imaging. First, mature leaves and shoots spectra were extracted from tea cutting seedlings using Mask R-CNN (Epoch=20 and Learning-rate=0.001). The spectra were then preprocessed by MSC, S-G, 1-D filtering techniques, with feature bands screened by UVE (Optimal factor number=5), CARS (Monte Carlo sampling times=300), and SPA (Epoch=25). Finally, a CNN-GRU (Epoch=100 and Learning-rate=0.01) network was employed to predict shoot and root biomass, and compared with machine learning methods of SVR (Kernel function=Polynomial), RFR (Ntrees=200), and PLSR (Latent variable=16) and deep learning methods (Epoch=100 and Learning-rate=0.01) of CNN and LSTM. The results show that (i) Mask R-CNN can accurately extract the spectral of mature leaves (precision=97.8 %) and shoots (precision=91.5 %), (ii) The spectral feature bands of shoot (number=212) and root (number=105) biomass screened by UVE were more abundant than CARS and SPA, (iii) The UVE+CNN-GRU model (Rp2=0.90, RMSEP = 0.12, RPD=2.43) based on shoots spectrum provided the optimal estimation of shoot biomass. The SPA+LSTM model (Rp2 = 0.65, RMSEP = 0.05, RPD=1.67) based on the mature leaves spectrum provided the optimal estimation of root biomass. These findings demonstrate that hyperspectral imaging technology combined with deep learning algorithms can accurately monitor the growth of cutting seedlings quickly without causing damage. This not only provides a new data sources along with technical means for efficient screening excellent tea varieties, but also improves agricultural production efficiency and resource utilization.

Detection of rice (with husk) moisture content based on hyperspectral imaging technology combined with MSLPP–ESMA–SVR model

AbstractMoisture content detection has guiding significance for the storage and quality detection of rice. To detect moisture content rapidly and non‐destructively, hyperspectral imaging technology (400‐1000 nm) was employed to analyze rice with different moisture content, and Savitzky–Golay mixed standard normalized variable algorithm (SG‐SNV) was used for spectral data pretreatment. Furthermore, a modified supervised locality preserving projections (MSLPP) method was proposed to extract spectral features. The modeling results showed that MSLPP had better spectral feature extraction performance. Finally, to improve prediction accuracy, the equilibrium slime mold algorithm (ESMA) was introduced to obtain the optimal parameters (c, g) of the support vector regression (SVR) model. And MSLPP–ESMA–SVR model had higher prediction accuracy and stronger robustness, with R2p reaching 0.9755 and root mean square error of prediction reaching 0.8597%. Therefore, hyperspectral imaging technology combined with MSLPP–ESMA–SVR model is feasible to detect rice moisture content.