Near-infrared Hyperspectral Imaging System Research Articles

Stand-off hazardous materials identification based on near-infrared hyperspectral imaging combined with convolutional neural network

Near-infrared (NIR) hyperspectral imaging enables rapid, non-contact imaging of hazardous materials in a non-destructive manner, allowing for analysis based on spectral reflection information. However, using traditional methods, it is challenging to identify hazardous materials with less distinct spectral reflection features. This study utilizes a self-built NIR hyperspectral imaging system and proposes a new approach. Using a convolutional neural network (CNN), This allows for the rapid completion of high-throughput spectral screening, marking suspicious spectra at spatial points. we sophisticatedly classified six hazardous material types, generating impactful warning images. The optimized CNN demonstrated superior performance (accuracy 91.08 %, recall 91.15 %, specificity 91.62 %, precision 90.17 %, and 0.924 F1 score) compared to SVM and KNN methods. Our study included multitask validation tests, revealing a sensitive detection of 10 mg/cm2 for ammonium nitrate and trinitrotoluene, capable of identifying over 100 targets simultaneously. By simulating real-world scenarios, we successfully detected hazardous chemicals scattered on the ground and accurately identified these hazardous materials in glass and thin plastic products. Even in situations where clothing obstructed the view, we could still correctly identify hazardous chemicals and generate corresponding warning images. Our system demonstrated precise identification capabilities even amidst complex backgrounds. This method provides an accurate and rapid solution for identifying and locating hazardous chemicals, laying a strong foundation for the next steps in non-contact, long-distance quantitative determination of chemical concentrations. This study highlights the effective application of CNN in non-contact, long-distance classification, and recognition of hazardous materials, paving the way for further scientific and engineering applications.

Simultaneous detection for storage condition and storage time of yellow peach under different storage conditions using hyperspectral imaging with multi-target characteristic selection and multi-task model

The detection for storage condition and storage time can ensure quality and safety of yellow peaches, and optimize storage management. This study explored the feasibility of simultaneously detecting the storage condition and storage time of yellow peach under different storage conditions using hyperspectral imaging technology combined with multi-target characteristic selection and multi-task model. Firstly, a total of 1080 hyperspectral images of 120 yellow peach samples under different storage conditions and different storage time were acquired using visible near-infrared hyperspectral imaging system. Subsequently, Savitzky-Golay first derivative was used to preprocess the raw spectra to improve the signal-to-noise ratio of the spectra. Distinguished index modified competitive adaptive reweighted sampling (DI-CARS) was proposed to select characteristic wavelengths that simultaneously characterize information of storage condition and storage time. Wild horse optimizer optimized support vector machine (WHO-SVM) was proposed for the multi-task modeling analysis. The results showed that multi-target characteristic selection and multi-task model not only reduced 66.67 % in computational cost, but also improved the accuracy, and multi-task WHO-SVM model based on multi-target DI-CARS achieved the optimal result with the overall accuracy reached 98.06 % and the number of characteristic wavelengths being 99. In conclusion, hyperspectral imaging combined with multi-target DI-CARS and multi-task WHO-SVM is feasible to simultaneously detect storage condition and storage time of yellow peach under different storage conditions.

Rapid determination of residual pefloxacin in mutton based on hyperspectral imaging and data fusion

Non-destructive detection of veterinary drug residues in meat products is critical for ensuring food safety and human health. The feasibility of using visible near-infrared (Vis-NIR) and near-infrared (NIR) hyperspectral imaging (HSI) system combined with data fusion for the prediction of pefloxacin residues in mutton was investigated in this study. A partial least squares regression optimization model was developed by multivariate data processing methods. The results showed that the low-level fusion (LLF) produced better results with R2P = 0.907 and RMSEP = 0.462 as compared to individual data blocks. Furthermore, intermediate-level fusion (ILF) based on iterative retention of informative variables algorithm showed the best results with R2P = 0.940 and RMSEP = 0.375. Finally, the visualized distributions of Vis-NIR and NIR based on the optimal model combination were mapped. The results have demonstrated that a combination of HSI with data fusion may be applied to rapidly and non-destructively detect pefloxacin residues in mutton.

Hyperspectral imaging combined with deep learning for discrimination of Lingwu long jujube in terms of the time after bruising

The timely detection and classification of bruised Lingwu long jujube is crucial for reducing economic losses in the agricultural industry. Traditional machine learning methods had been effectively used for bruise time classification with limited samples while the challenges were presented when the amount of data expanded. Deep learning methods can offer a solution. The hyperspectral image of the jujube sample was obtained using near-infrared hyperspectral imaging system (900–1700 nm). The optimal input variables were obtained by comparing the performance of the partial least squares-discriminant analysis, artificial neural network, and one-dimensional convolutional neural network (1D-CNN) models based on the spectral data and the fusion data of spectral and textural features. The optimal variable was the fusion data that could improve the model performance (from 92.68% to 96.10%). The above-mentioned three classifiers were used to establish models based on fusion data for two different dataset sizes consisting of large datasets (LD) (820) and small datasets (SD) (200) to explore the influence of dataset size on model performance. As the sample size increased, the 1D-CNN had good robustness with a classification rate of 94.00% for SD and 96.10% for LD in the prediction sets for all samples. The classification rate of prediction sets for 0 h after bruising increased from 90.00% to 92.68%, and from 80.00% to 90.24% in the class of 24 h after bruising. The findings provided a new approach to building a large spectral library and developing an online classification and detection system, thereby providing new ideas for the development of grading systems for Lingwu long jujube.

Improving the performance of the model developed from the classification of adulterated honey with different botanical origins based on near-Infrared hyperspectral imaging system and supervised classification algorithms

Honey is a high-value product that proscribes adulteration. Authentication of pure honey is an imperative act that can disregard the adverse impacts brought by fraudsters on honey producers and consumers. A reliable method that can detect food adulteration that is nondestructive, rapid, and efficient can help maintain product authenticity. This study investigated the potential of a near-infrared hyperspectral imaging (NIR-HSI) system combined with supervised classification techniques for the classification of honey samples from different botanical origins that were adulterated with sugar cane syrup at 20 % and 40 % w/w. Reflectance and transflectance modes of measurement were used to obtain honey spectral images from 864.53 nm to 1700 nm. Supervised classification algorithms, namely, stepwise discriminant analysis (SDA), support vector machine (SVM), and k-nearest neighbors (kNN) were used for the classification based on whole spectral data and feature selected spectral data. The best classification model from reflectance mode was observed from SDA based on a leave-one-out cross-validation (LOOCV) method with a test accuracy of 95 %. For transflectance mode, the classification model developed from SDA based on a LOOCV method provided better test set accuracy of 100 %. The classification performance levels of the models developed from SVM and kNN were lower with the use of the whole wavelength region. Moreover, the classification performance levels of SVM and kNN using the selected features by the SDA greatly increased, however, they were not as good as from using SDA alone. The classifier maps created using SDA and predicting the averaged spectra of all pixels in the image with 10-discriminant functions for each mode of measurement displayed a two-dimensional image that had a distinct colour pertaining to a particular class that could easily discriminate one class from another.

Near-infrared hyperspectral imaging combined with machine learning for physicochemical-based quality evaluation of durian pulp

This research reports on the application of near-infrared hyperspectral imaging (NIR-HSI) system for predicting the physicochemical properties; dry matter (DM), total soluble solids (TSS), and fat content (FC) of durian. Partial least squares regression (PLSR), support vector machine (SVM), random forest (RF), and 1D convolution neural network (CNN) models: custom, U-Net, and VGG19; were developed to predict DM, TSS, and FC of durian pulp. Feature wavelengths were selected using a genetic algorithm (GA) and successive projection algorithm (SPA). The selected wavelengths were then validated based on the algorithms for regression model development. GA-PLSR model was compelling to predict the DM and FC in durian pulp, which obtained the coefficient of determination for the test set (r2) and root mean square error of prediction (RMSEP) of 0.97 and 1.12% for DM and 0.86 and 0.64% for FC, respectively. The GA-PLSR model provided the best result for the TSS prediction with r2, and RMSEP of 0.90 and 1.40%, respectively, whereas the SPA-PLSR model based on only thirteen wavelengths attained fair result with the r2 and RMSEP of 0.79 and 2.03%, respectively. The above results show that the pushbroom NIR-HSI system achieved promising results for estimating DM, TSS, and FC in durian pulp. This research identified the featured wavelengths that can be used to develop a portable and reliable HSI or multispectral system to be installed at durian packaging firms for quality inspection and grading.

Rapid shipboard measurement of net-collected marine microplastic polymer types using near-infrared hyperspectral imaging

Isolation and detection of microplastics (MP) in marine samples is extremely cost- and labor-intensive, limiting the speed and amount of data that can be collected. In the current work, we describe rapid measurement of net-collected MPs (net mesh size 300 µm) using a benchtop near-infrared hyperspectral imaging system during a research expedition to the subtropical North Atlantic gyre. Suspected plastic particles were identified microscopically and mounted on a black adhesive background. Particles were imaged with a Specim FX17 near-infrared linescan camera and a motorized stage. A particle mapping procedure was built on existing edge-finding algorithms and a polymer identification method developed using spectra from virgin polymer reference materials. This preliminary work focused on polyethylene, polypropylene, and polystyrene as they are less dense than seawater and therefore likely to be found floating in the open ocean. A total of 27 net tows sampled 2534 suspected MP particles that were imaged and analyzed at sea. Approximately 77.1% of particles were identified as polyethylene, followed by polypropylene (9.2%). A small fraction of polystyrene was detected only at one station. Approximately 13.6% of particles were either other plastic polymers or were natural materials visually misidentified as plastics. Particle size distributions for PE and PP particles with a length greater than 1 mm followed an approximate power law relationship with abundance. This method allowed at-sea, near real-time identification of MP polymer types and particle dimensions, and shows great promise for rapid field measurements of microplastics in net-collected samples.

Single Seed Near-Infrared Hyperspectral Imaging for Classification of Perennial Ryegrass Seed.

The detection of beneficial microbes living within perennial ryegrass seed causing no apparent defects is challenging, even with the most sensitive and conventional methods, such as DNA genotyping. Using a near-infrared hyperspectral imaging system (NIR-HSI), we were able to discriminate not only the presence of the commercial NEA12 fungal endophyte strain but perennial ryegrass cultivars of diverse seed age and batch. A total of 288 wavebands were extracted for individual seeds from hyperspectral images. The optimal pre-processing methods investigated yielded the best partial least squares discriminant analysis (PLS-DA) classification model to discriminate NEA12 and without endophyte (WE) perennial ryegrass seed with a classification accuracy of 89%. Effective wavelength (EW) selection based on GA-PLS-DA resulted in the selection of 75 wavebands yielding 88.3% discrimination accuracy using PLS-DA. For cultivar identification, the artificial neural network discriminant analysis (ANN-DA) was the best-performing classification model, resulting in >90% classification accuracy for Trojan, Alto, Rohan, Governor and Bronsyn. EW selection using GA-PLS-DA resulted in 87 wavebands, and the PLS-DA model performed the best, with no extensive compromise in performance, resulting in >89.1% accuracy. The study demonstrates the use of NIR-HSI reflectance data to discriminate, for the first time, an associated beneficial fungal endophyte and five cultivars of perennial ryegrass seed, irrespective of seed age and batch. Furthermore, the negligible effects on the classification errors using EW selection improve the capability and deployment of optimized methods for real-time analysis, such as the use of low-cost multispectral sensors for single seed analysis and automated seed sorting devices.

Development of a General Prediction Model of Moisture Content in Maize Seeds Based on LW-NIR Hyperspectral Imaging

Moisture content (MC) is one of the important indexes to evaluate maize seed quality. Its accurate prediction is very challenging. In this study, the long-wave near-infrared hyperspectral imaging (LW-NIR-HSI) system was used, and the embryo side (S1) and endosperm side (S2) spectra of each maize seed were extracted, as well as the average spectrum (S3) of both being calculated. The partial least square regression (PLSR) and least-squares support vector machine (LS-SVM) models were established. The uninformative variable elimination (UVE) and successive projections algorithm (SPA) were employed to reduce the complexity of the models. The results indicated that the S3-UVE-SPA-PLSR and S3-UVE-SPA-LS-SVM models achieved the best prediction accuracy with an RMSEP of 1.22% and 1.20%, respectively. Furthermore, the combination (S1+S2) of S1 and S2 was also used to establish the prediction models to obtain a general model. The results indicated that the S1+S2-UVE-SPA-LS-SVM model was more valuable with Rpre of 0.91 and RMSEP of 1.32% for MC prediction. This model can decrease the influence of different input spectra (i.e., S1 or S2) on prediction performance. The overall study indicated that LW-HSI technology combined with the general model could realize the non-destructive and stable prediction of MC in maize seeds.

Raw Beef Patty Analysis Using Near-Infrared Hyperspectral Imaging: Identification of Four Patty Categories

South African legislation regulates the classification/labelling and compositional specifications of raw beef patties, to combat processed meat fraud and to protect the consumer. A near-infrared hyperspectral imaging (NIR-HSI) system was investigated as an alternative authentication technique to the current destructive, time-consuming, labour-intensive and expensive methods. Eight hundred beef patties (ca. 100 g) were made and analysed to assess the potential of NIR-HSI to distinguish between the four patty categories (200 patties per category): premium 'ground patty'; regular 'burger patty'; 'value-burger/patty' and the 'econo-burger'/'budget'. Hyperspectral images were acquired with a HySpex SWIR-384 (short-wave infrared) imaging system using the Breeze® acquisition software, in the wavelength range of 952-2517 nm, after which the data was analysed using image analysis, multivariate techniques and machine learning algorithms. It was possible to distinguish between the four patty categories with accuracies ≥97%, indicating that NIR-HSI offers an accurate and reliable solution for the rapid identification and authentication of processed beef patties. Furthermore, this study has the potential of providing an alternative to the current authentication methods, thus contributing to the authenticity and fair-trade of processed meat products locally and internationally.

Near-infrared hyperspectral imaging for online measurement of the viability detection of naturally aged watermelon seeds.

The viability status of seeds before sowing is important to farmers as it allows them to make yield predictions. Monitoring the seed quality in a rapid and nondestructive manner may create a perfect solution, especially for industrial sorting applications. However, current offline laboratory-based strategies employed for the monitoring of seed viability are time-consuming and thus cannot satisfy industrial needs where there is a substantial number of seeds to be analyzed. In this study, we describe a prototype online near-infrared (NIR) hyperspectral imaging system that can be used for the rapid detection of seed viability. A wavelength range of 900-1700 nm was employed to obtain spectral images of three different varieties of naturally aged watermelon seed samples. The partial least square discriminant analysis (PLS-DA) model was employed for real-time viability prediction for seed samples moving through a conveyor unit at a speed of 49 mm/sec. A suction unit was further incorporated to develop the online system and it was programmatically controlled to separate the detected viable seeds from nonviable ones. For an external validation sample set showed classification accuracy levels of 91.8%, 80.7%, and 77.8% in relation to viability for the three varieties of watermelon seed with healthy seedling growth. The regression coefficients of the classification model distinguished some chemical differences in viable and nonviable seed which was verified by the chromatographic analysis after the detection of the proposed online system. The results demonstrated that the developed online system with the viability prediction model has the potential to be used in the seed industry for the quality monitoring of seeds.

Hyperspectral imaging solutions for the non-invasive detection and automated mapping of copper trihydroxychlorides in ancient bronze

Ancient bronze is subject to complex degradation which can lead, in cases where copper chlorides are present, to a cyclic and self-sustaining degradation process commonly referred to as “bronze disease”. If left untreated, bronze disease can eat away at a bronze object until it is entirely deteriorated. The presence of copper trihydroxychlorides is indicative that this process is underway and therefore the detection of these corrosion products is necessary in guiding conservation of ancient bronze artefacts. In this paper we present a high spatial/spectral resolution short wave infrared (SWIR) imaging solution for mapping copper trihydroxychlorides in ancient bronze, combining hyperspectral imaging with an in-house developed unsupervised machine learning algorithm for automated spectral clustering. For this work, verification was obtained through use of an in-house developed reference database of typical ancient bronze corrosion products from several archaeological sites, and from collections of the National Museum of China. This paper also explores the suitability, and limitations, of a visible to near-infrared (VNIR) hyperspectral imaging system as a more accessible solution for mapping copper trihydroxychlorides associated with bronze disease. We suggest that our hyperspectral imaging solution can provide a non-invasive, rapid, and high resolution material mapping within and across bronze objects, particularly beneficial for analysing large collections in a museum setting.

Finite Element Analysis and Near-Infrared Hyperspectral Reflectance Imaging for the Determination of Blueberry Bruise Grading.

Bruising of the subcutaneous tissues of blueberries is an important form of mechanical damage. Different levels of bruising have a significant effect on the post-harvest marketing of blueberries. To distinguish different grades of blueberry bruises and explore the effects of different factors, explicit dynamic simulation and near-infrared hyperspectral reflectance imaging were employed without harming the blueberries in this study. Based on the results of the compression experiment, an explicit dynamic simulation of blueberries was performed to measure the potential locations of bruises and preliminarily divide the bruise stages. A near-infrared hyperspectral reflectance imaging system was used to detect the actual blueberry bruises. According to the blueberry photos taken by the near-infrared hyperspectral reflectance imaging system, the actual bruise rates of blueberries were obtained by using the Environment for Visualizing Images software for training and classification. Bruise grades of blueberries were divided accordingly. Response surface methodology was used to determine the effects of ripeness, loading speed and loading location on the blueberry bruising rate. Under the optimized parameters, the actual damage rate of blueberries was 1.1%. The results provide an important theoretical basis for the accurate and rapid identification and classification of blueberry bruise damage.

Exploring the potential of hyperspectral imaging to detect Esca disease complex in asymptomatic grapevine leaves

• A pixel-based classification of esca-affected grapevine leaves was achieved by HSI. • Multiple pre-processing and variable selection methods were tested using PLS-DA. • Feature wavelengths were identified using automatic and interval PLS methods. • Asymptomatic leaves belonging to esca-affected vines were successfully identified. • NIR-HSI could detect foliar symptoms of grapevine Esca disease complex. Precise and reliable identification of specific plant diseases is a challenge within precision agriculture nowadays. This is the case of esca, a complex grapevine trunk disease, that represents a major threat to modern viticulture as it is responsible for large economic losses annually. The lack of effective control strategies and the complexity of esca disease expression make essential the identification of affected plants, before symptoms become evident, for a better management of the vineyard. This study evaluated the suitability of a near-infrared hyperspectral imaging (HSI) system to detect esca disease in asymptomatic grapevine leaves of Tempranillo red-berried cultivar. For this, 72 leaves from an experimental vineyard, naturally infected with esca, were collected and scanned with a lab-scale HSI system in the 900-1700 nm spectral range. Then, effective image processing and multivariate analysis techniques were merged to develop pixel-based classification models for the distinction of healthy, asymptomatic and symptomatic leaves. Automatic and interval partial least squares variable selection methods were tested to identify the most relevant wavelengths for the detection of esca-affected vines using partial least squares discriminant analysis and different pre-processing techniques. Three-class and two-class classifiers were carried out to differentiate healthy, asymptomatic and symptomatic leaf pixels, and healthy from asymptomatic pixels, respectively. Both variable selection methods performed similarly, achieving good classification rates in the range of 82.77-97.17% in validation datasets for either three-class or two-class classifiers. The latter results demonstrated the capability of hyperspectral imaging to distinguish two groups of seemingly identical leaves (healthy and asymptomatic). These findings would ease the annual monitoring of disease incidence in the vineyard and, therefore, better crop management and decision making.

Predicting the true density of commercial biomass pellets using near-infrared hyperspectral imaging

The use of biomass is increasing because it is a form of renewable energy that provides high heating value. Rapid measurements could be used to check the quality of biomass pellets during production. This research aims to apply a near-infrared (NIR) hyperspectral imaging system for the evaluation of the true density of individual biomass pellets during the production process. Real-time measurement of the true density could be beneficial for the operation settings, such as the ratio of the binding agent to the raw material, the temperature of operation, the production rate, and the mixing ratio. The true density could also be used for rough measurement of the bulk density, which is a necessary parameter in commercial production. Therefore, knowledge of the true density is required during production in order to maintain the pellet quality as well as operation conditions. A prediction model was developed using partial least squares (PLS) regression across different wavelengths selected using different spectral pre-treatment methods and variable selection methods. After model development, the performance of the models was compared. The best model for predicting the true density of individual pellets was developed with first-derivative spectra (D1) and variables selected by the genetic algorithm (GA) method, and the number of variables was reduced from 256 to 53 wavelengths. The model gave R2cal, R2val, SEC, SEP, and RPD values of 0.88, 0.89, 0.08 g/cm3, 0.07 g/cm3, and 3.04, respectively. The optimal prediction model was applied to construct distribution maps of the true density of individual biomass pellets, with the level of the predicted values displayed in colour bars. This imaging technique could be used to check visually the true density of biomass pellets during the production process for warnings to quality control equipment.

Near-Infrared Hyperspectral Imaging as a Monitoring Tool for On-Demand Manufacturing of Inkjet-Printed Formulations

This study evaluates the potential use of near-infrared hyperspectral imaging (NIR-HSI) for quantitative determination of the drug amount in inkjet-printed dosage forms. We chose metformin hydrochloride as a model active pharmaceutical ingredient (API) and printed it onto gelatin films using a piezoelectric inkjet printing system. An industry-ready NIR-HSI sensor combined with a motorized movable linear stage was applied for spectral acquisition. Initial API-substrate screening revealed best printing results for gelatin films with TiO2 filling. For calibration of the NIR-HSI system, escalating drug doses were printed on the substrate. After spectral pre-treatments, including standard normal variate (SNV) and Savitzky-Golay filtering for noise reduction and enhancement of spectral features, principal component analysis (PCA) and partial least squares (PLS) regression were applied to create predictive models for the quantification of independent printed metformin hydrochloride samples. It could be shown that the concentration distribution maps provided by the developed HSI models were capable of clustering and predicting the drug dose in the formulations. HSI model prediction showed significant better correlation to the reference (HPLC) compared to on-board monitoring of dispensed volume of the printer. Overall, the results emphasize the capability of NIR-HSI as a fast and non-destructive method for the quantification and quality control of the deposited API in drug-printing applications.Graphical abstract

Near-Infrared Hyperspectral Imaging (NIR-HSI) for Nondestructive Prediction of Anthocyanins Content in Black Rice Seeds

Anthocyanins are an important micro-component that contributes to the quality factors and health benefits of black rice. Anthocyanins concentration and compositions differ among rice seeds depending on the varieties, growth conditions, and maturity level at harvesting. Chemical composition-based seeds inspection on a real-time, non-destructive, and accurate basis is essential to establish industries to optimize the cost and quality of the product. Therefore, this research aimed to evaluate the feasibility of near-infrared hyperspectral imaging (NIR-HSI) to predict the content of anthocyanins in black rice seeds, which will open up the possibility to develop a sorting machine based on rice micro-components. Images of thirty-two samples of black rice seeds, harvested in 2019 and 2020, were captured using the NIR-HSI system with a wavelength of 895–2504 nm. The spectral data extracted from the image were then synchronized with the rice anthocyanins reference value analyzed using high-performance liquid chromatography (HPLC). For comparison, the seed samples were ground into powder, which was also captured using the same NIR-HSI system to obtain the data and was then analyzed using the same method. The model performance of partial least square regression (PLSR) of the seed sample developed based on harvesting time, and mixed data revealed the model consistency with R2 over 0.85 for calibration datasets. The best prediction models for 2019, 2020, and mixed data were obtained by applying standard normal variate (SNV) pre-processing, indicated by the highest coefficient of determination (R2) of 0.85, 0.95, 0.90, and the lowest standard error of prediction (SEP) of 0.11, 0.17, and 0.16 mg/g, respectively. The obtained R2 and SEP values of the seed model were comparable to the result of powder of 0.92–0.95 and 0.09–0.15 mg/g, respectively. Additionally, the obtained beta coefficients from the developed model were used to generate seed chemical images for predicting anthocyanins in rice seed. The root mean square error (RMSE) value for seed prediction evaluation showed an acceptable result of 0.21 mg/g. This result exhibits the potential of NIR-HSI to be applied in a seed sorting machine based on the anthocyanins content.

Artificial Intelligence Empowered Multispectral Vision Based System for Non-Contact Monitoring of Large Yellow Croaker (Larimichthys crocea) Fillets.

A non-contact method was proposed to monitor the freshness (based on TVB-N and TBA values) of large yellow croaker fillets (Larimichthys crocea) by using a visible and near-infrared hyperspectral imaging system (400–1000 nm). In this work, the quantitative calibration models were built by using feed-forward neural networks (FNN) and partial least squares regression (PLSR). In addition, it was established that using a regression coefficient on the data can be further compressed by selecting optimal wavelengths (35 for TVB-N and 18 for TBA). The results validated that FNN has higher prediction accuracies than PLSR for both cases using full and selected reflectance spectra. Moreover, our FNN based model has showcased excellent performance even with selected reflectance spectra with rp = 0.978, R2p = 0.981, and RMSEP = 2.292 for TVB-N, and rp = 0.957, R2p = 0.916, and RMSEP = 0.341 for TBA, respectively. This optimal FNN model was then utilized for pixel-wise visualization maps of TVB-N and TBA contents in fillets.

Real-time determination and visualization of two independent quantities during a manufacturing process of pharmaceutical tablets by near-infrared hyperspectral imaging combined with multivariate analysis.

During pharmaceutical manufacturing, line-scan hyperspectral imaging enables us to collect several electromagnetic spectra at each pixel in a two-dimensional plane for each tablet. The present study quantitatively determines two independent values of the active pharmaceutical ingredient (API) content in a tablet and the amount of coating on a surface of the same tablet simultaneously; the process is visualized by means of a near-infrared hyperspectral imaging (NIR-HSI) system combined with multivariate data analysis at a typical manufacturing speed of 4,000 tablets per minute. The API content and the amount of coating were controlled to be in the range 80-120% and 0-7mg, respectively. The results of the cross validation of regression models demonstrated a coefficient of determination (R2) of 0.942, a root-mean-square error of cross validation (RMSECV) of 3.48% for the API content, an R2 of 0.939, and an RMSECV of 0.46mg for the amount of coating. These results demonstrated that the API content in a tablet as well as the amount of coating on the surface of the same tablet can be simultaneously determined with sufficient accuracy. This technique is practically applicable to process analytical technology in pharmaceutical manufacturing.

Online Application of a Hyperspectral Imaging System for the Sorting of Adulterated Almonds

Almonds are nutrient-rich nuts. Due to their high level of consumption and relatively high price, their production is targeted for illegal practices, with the intention of earning more profit. The most common adulterants are based on superficial matching, and as an adulterant, the apricot kernel is comparatively inexpensive and almost identical in color, texture, odor, and other physicochemical characteristics to almonds. In this study, a near-infrared hyperspectral imaging (NIR-HSI) system in the wavelength range of 900–1700 nm synchronized with a conveyor belt was used for the online detection of added apricot kernels in almonds. A total of 448 samples from different varieties of almonds and apricot kernels (112 × 4) were scanned while the samples moved on the conveyor belt. The spectral data were extracted from each imaged nut and used to develop a partial least square discrimination analysis (PLS-DA) model coupled with different preprocessing techniques. The PLS-DA model displayed over a 97% accuracy for the validation set. Additionally, the beta coefficient obtained from the developed model was used for pixel-based classification. An image processing algorithm was developed for the chemical mapping of almonds and apricot kernels. Consequently, the obtained model was transferred for the online sorting of seeds. The online classification system feedback had an overall accuracy of 85% for the classification of nuts. However, the model presented a relatively low accuracy when evaluated in real-time for online application, which might be due to the rough distribution of samples on the conveyor belt, high speed, delaying time in suction, and lighting variations. Nevertheless, the developed online prototype (NIR-HSI) system combined with multivariate analysis exhibits strong potential for the classification of adulterated almonds, and the results indicate that the system can be effectively used for the high-throughput screening of adulterated almond nuts in an industrial environment.