Near Infrared Spectroscopy (NIRS) for the determination of the milk fat fatty acid profile of goats

Visible/near-infrared hyperspectral imaging combined with machine learning for identification of ten Dalbergia species

Introduction

This study addresses the urgent need for non-destructive identification of commercially valuable Dalbergia species, which are threatened by illegal logging. Effective identification methods are crucial for ecological conservation, biodiversity preservation, and the regulation of the timber trade.

Methods

We integrate Visible/Near-Infrared (Vis/NIR) Hyperspectral Imaging (HSI) with advanced machine learning techniques to enhance the precision and efficiency of wood species identification. Our methodology employs various modeling approaches, including Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA), Support Vector Machine (SVM), and Convolutional Neural Networks (CNN). These models analyze spectral data across Vis (383-982 nm), NIR (982-2386 nm), and full spectral ranges (383 nm to 2386 nm). We also assess the impact of preprocessing techniques such as Standard Normal Variate (SNV), Savitzky-Golay (SG) smoothing, normalization, and Multiplicative Scatter Correction (MSC) on model performance.

Results

With optimal preprocessing, both SVM and CNN models achieve 100% accuracy across NIR and full spectral ranges. The selection of an appropriate wavelength range is critical; utilizing the full spectrum captures a broader array of the wood's chemical and physical properties, significantly enhancing model accuracy and predictive power.

Discussion

These findings underscore the effectiveness of Vis/NIR HSI in wood species identification. They also highlight the importance of precise wavelength selection and preprocessing techniques to maximize both accuracy and cost-efficiency. This research contributes substantially to ecological conservation and the regulation of the timber trade by providing a reliable, non-destructive method for identifying threatened wood species.

Machine Learning-Enabled NIR Spectroscopy. Part 3: Hyperparameter by Design (HyD) Based ANN-MLP Optimization, Model Generalizability, and Model Transferability

Data variations, library changes, and poorly tuned hyperparameters can cause failures in data-driven modelling. In such scenarios, model drift, a gradual shift in model performance, can lead to inaccurate predictions. Monitoring and mitigating drift are vital to maintain model effectiveness. USFDA and ICH regulate pharmaceutical variation with scientific risk-based approaches. In this study, the hyperparameter optimization for the Artificial Neural Network Multilayer Perceptron (ANN-MLP) was investigated using open-source data. The design of experiments (DoE) approach in combination with target drift prediction and statistical process control (SPC) was employed to achieve this objective. First, pre-screening and optimization DoEs were conducted on lab-scale data, serving as internal validation data, to identify the design space and control space. The regression performance metrics were carefully monitored to ensure the right set of hyperparameters was selected, optimizing the modelling time and storage requirements. Before extending the analysis to external validation data, a drift analysis on the target variable was performed. This aimed to determine if the external data fell within the studied range or required retraining of the model. Although a drift was observed, the external data remained well within the range of the internal validation data. Subsequently, trend analysis and process monitoring for the mean absolute error of the active content were conducted. The combined use of DoE, drift analysis, and SPC enabled trend analysis, ensuring that both current and external validation data met acceptance criteria. Out-of-specification and process control limits were determined, providing valuable insights into the model's performance and overall reliability. This comprehensive approach allowed for robust hyperparameter optimization and effective management of model lifecycle, crucial in achieving accurate and dependable predictions in various real-world applications.

Discrimination of Milk Freshness Based on Synchronous Two-Dimensional Visible/Near-Infrared Correlation Spectroscopy Coupled with Chemometrics

Milk is one of the preferred beverages in modern healthy diets, and its freshness is of great significance for product sales and applications. By combining the two-dimensional (2D) correlation spectroscopy technique and chemometrics, a new method based on visible/near-infrared (Vis/NIR) spectroscopy was proposed to discriminate the freshness of milk. To clarify the relationship be-tween the freshness of milk and the spectra, the changes in the physicochemical indicators of milk during storage were analyzed as well as the Vis/NIR spectra and the 2D-Vis/NIR correlation spectra. The threshold-value method, linear discriminant analysis (LDA) method, and support vector machine (SVM) method were used to construct the discriminant models of milk freshness, and the parameters of the SVM-based models were optimized by the grid search method and particle swarm optimization algorithm. The results showed that with the prolongation of storage time, the absorbance of the Vis/NIR spectra of milk gradually increased, and the intensity of autocorrelation peaks and cross peaks in synchronous 2D-Vis/NIR spectra also increased significantly. Compared with the SVM-based models using Vis/NIR spectra, the SVM-based model using 2D-Vis/NIR spectra had a >15% higher prediction accuracy. Under the same conditions, the prediction performances of the SVM-based models were better than those of the threshold-value-based or LDA-based models. In addition, the accuracy rate of the SVM-based model using the synchronous 2D-Vis/NIR autocorrelation spectra was >97%. This work indicates that the 2D-Vis/NIR correlation spectra coupled with chemometrics is a great pattern to rapidly discriminate the freshness of milk, which provides technical support for improving the evaluation system of milk quality and maintaining the safety of milk product quality.

Phenotypic Analysis of Fourier-Transform Infrared Milk Spectra in Dairy Goats

The infrared spectrum of bovine milk is used to predict many interesting traits, whereas there have been few studies on goat milk in this regard. The objective of this study was to characterize the major sources of variation in the absorbance of the infrared spectrum in caprine milk samples. A total of 657 goats belonging to 6 breeds and reared on 20 farms under traditional and modern dairy systems were milk-sampled once. Fourier-transform infrared (FTIR) spectra were taken (2 replicates per sample, 1314 spectra), and each spectrum contained absorbance values at 1060 different wavenumbers (5000 to 930 × cm^-1), which were treated as a response variable and analyzed one at a time (i.e., 1060 runs). A mixed model, including the random effects of sample/goat, breed, flock, parity, stage of lactation, and the residual, was used. The pattern and variability of the FTIR spectrum of caprine milk was similar to those of bovine milk. The major sources of variation in the entire spectrum were as follows: sample/goat (33% of the total variance); flock (21%); breed (15%); lactation stage (11%); parity (9%); and the residual unexplained variation (10%). The entire spectrum was segmented into five relatively homogeneous regions. Two of them exhibited very large variations, especially the residual variation. These regions are known to be affected by the absorbance of water, although they also exhibited wide variations in the other sources of variation. The average repeatability of these two regions were 45% and 75%, whereas for the other three regions it was about 99%. The FTIR spectrum of caprine milk could probably be used to predict several traits and to authenticate the origin of goat milk.

New research development on trans fatty acids in food: Biological effects, analytical methods, formation mechanism, and mitigating measures

The trans fatty acids (TFAs) in food are mainly generated from the ruminant animals (meat and milk) and processed oil or oil products. Excessive intake of TFAs (>1% of total energy intake) caused more than 500,000 deaths from coronary heart disease and increased heart disease risk by 21% and mortality by 28% around the world annually, which will be eliminated in industrially-produced trans fat from the global food supply by 2023. Herein, we aim to provide a comprehensive overview of the biological effects, analytical methods, formation and mitigation measures of TFAs in food. Especially, the research progress on the rapid, easy-to-use, and newly validated analytical methods, new formation mechanism, kinetics, possible mitigation mechanism, and new or improved mitigation measures are highlighted. We also offer perspectives on the challenges, opportunities, and new directions for future development, which will contribute to the advances in TFAs research.

Determination of nutritional health indexes of fresh bovine milk using near infrared spectroscopy

Bovine milk is one of the most complete foods that exist. During the last decades, milk FA have shown to improve human health due to the reduction in risk of cardiovascular disease and related pathologies. The aim of this study was to evaluate the feasibility of near infrared spectroscopy (NIRS) reflectance analysis to predict the nutritional value, fatty acid (FA) composition, and health index of fresh milk from dairy cows of pastoral systems. The prediction of Atherogenicity and Thrombogenicity indexes, along with other FA ratios in fresh milk samples by NIRS were precise and accurate. In addition, the calibration model obtained by NIRS provides an opportunity for the routine quantification of milk’s healthy FA such as omega-3 and conjugated linoleic acid (CLA), with applications in the dairy industry for food labeling, and at the farm level for management of the dairy cow’s diet.

Spectroscopic-Based Prediction of Milk Foam Properties for Barista Applications

The important quality parameters of cow’s milk for barista applications are frothability and foam stability. In the past, quality assessment was very time-consuming and could only be carried out after milk treatment had been completed. Since spectroscopy is already established in dairies, it could be advantageous to develop a spectrometer-based measurement method for quality control for barista applications. By integrating online spectroscopy to the processing of UHT (ultra-high temperature processing) milk before filling, it can be checked whether the currently processed product is suitable for barista applications. To test this hypothesis, a feasibility study was conducted. For this purpose, seasonal UHT whole milk samples were measured every 2 months over a period of more than 1 year, resulting in a total of 269 milk samples that were foamed. Samples were frothed using a self-designed laboratory frother. Frothability at the beginning and foam loss after 15 min describe the frothing characteristics of the milk and are predicted from the spectra. Near-infrared, Raman, and fluorescence spectra were recorded from each milk sample. These spectra were preprocessed using 15 different mathematical methods. For each spectrometer, 85% of the resulting spectral dataset was analyzed using partial least squares (PLS) regression and nine different variable selection (VS) algorithms. Using the remaining 15% of the spectral dataset, a prediction error was determined for each model and used to compare the models. Using spectroscopy and PLS modeling, the best results show a prediction error for milk frothability of 3% and foam stability of 2%.

Determination of trans-fatty acids in food samples based on the pre-column fluorescence derivatization by high performance liquid chromatography

Trans-fatty acids are unsaturated fatty acids that are considered to have health risks. 1,3,5,7-Tetramethyl-8-butyrethylenediamine-difluoroboradiaza-s-indacene is a highly-sensitive fluorescent labeling reagent for carboxylic acids developed by our lab. In this study, using this pre-column fluorescent derivatization reagent, a rapid and accurate high-performance liquid chromatography-fluorescence detection method was developed for the determination of two trans-fatty acids in food samples. Under the optimized derivative conditions, two trans-fatty acids were tagged with the fluorescent labeling reagent in the presence of 1-ethyl-3-(3-dimethyl-aminopropyl) carbodiimide at 25 °C for 30 min. Then, the baseline separation of trans- and cis-fatty acids and their saturated fatty acid with similar structures was achieved with less interference using a reversed-phased C₁₈ column with isocratic elution in 14 min. With fluorescence detection at λ_ex /λ_em = 490 nm/510 nm, the linear range of the trans-fatty acids was 1.0-200 nM with low detection limits in the range of 0.1-0.2 nM (signal-to-noise ratio = 3). In addition, the proposed approach was successfully applied for the detection of trans-fatty acids in food samples, and the recoveries using this method ranged from 96.02% to 109.22% with low relative standard deviations of 1.2-4.3% (n = 6). This article is protected by copyright. All rights reserved.

Application of Artificial Intelligence in Food Industry—a Guideline

Artificial intelligence (AI) has embodied the recent technology in the food industry over the past few decades due to the rising of food demands in line with the increasing of the world population. The capability of the said intelligent systems in various tasks such as food quality determination, control tools, classification of food, and prediction purposes has intensified their demand in the food industry. Therefore, this paper reviews those diverse applications in comparing their advantages, limitations, and formulations as a guideline for selecting the most appropriate methods in enhancing future AI- and food industry–related developments. Furthermore, the integration of this system with other devices such as electronic nose, electronic tongue, computer vision system, and near infrared spectroscopy (NIR) is also emphasized, all of which will benefit both the industry players and consumers.

Real-Time Monitoring of Yogurt Fermentation Process by Aquaphotomics Near-Infrared Spectroscopy

Automated quality control could have a substantial economic impact on the dairy industry. At present, monitoring of yogurt production is performed by sampling for microbiological and physicochemical measurements. In this study, Near-Infrared Spectroscopy (NIRS) is proposed for non-invasive automated control of yogurt production and better understanding of lactic acid bacteria (LAB) fermentation. UHT (ultra-high temperature) sterilized milk was inoculated with Bulgarian yogurt and placed into a quartz cuvette (1 mm pathlength) and test-tubes. Yogurt absorbance spectra (830-2500 nm) were acquired every 15 min, and pH, in the respective test-tubes, was measured every 30 min, during 8 h of fermentation. Spectral data showed substantial baseline and slope changes with acidification. These variations corresponded to respective features of the microbiological growth curve showing water structural changes, protein denaturation, and coagulation of milk. Moving Window Principal Component Analysis (MWPCA) was applied in the spectral range of 954-1880 nm to detect absorbance bands where most variations in the loading curves were caused by LAB fermentation. Characteristic wavelength regions related to the observed physical and multiple chemical changes were identified. The results proved that NIRS is a valuable tool for real-time monitoring and better understanding of the yogurt fermentation process.

Proximal Composition and Nutritive Value of Raw, Smoked and Pickled Freshwater Fish

The aim of the study was to assess protein, fat, salt, collagen, moisture content and energy value of freshwater fish purchased in Polish fish farms. Eight species of freshwater fish (raw, smoked, pickled) were assessed by near infrared spectroscopy (NIRS). The protein content varied between 15.9 and 21.7 g/100 g, 12.8 and 26.2 g/100 g, 11.5 and 21.9 g/100 g in raw, smoked and pickled fish, respectively. Fat content ranged from 0.89 to 22.3 g/100 g, 0.72 to 43.1 g/100 g, 0.01 to 29.7 g/100 g in raw, smoked and pickled fish, respectively. Salt content ranged from 0.73 to 1.48 g/100 g, 0.77 to 3.39 g/100 g, 1.47 to 2.29 g/100 g in raw, smoked and pickled fish, respectively. A serving (150 g) of each fish product provided 53.2–71.9% of the Reference Intake (RI) for protein, 2.21–60.3% of the RI for fat, 21.3–61.3% of the RI for salt and 6.27–24.4% kJ/6.29–24.5% kcal of the RI for energy. Smoked fish had a higher protein and also fat content than raw and pickled fish, while smoked and pickled fish had higher salt content than raw fish. Cluster analysis was performed, which allowed to distinguish, on the basis of protein, fat, salt, collagen and moisture content, mainly European eel.

The recent advances of near-infrared spectroscopy in dairy production-a review

One of the major issues confronting the dairy industry is the efficient evaluation of the quality of feed, milk and dairy products. Over the years, the use of rapid analytical methods in the dairy industry has become imperative. This is because of the documented evidence of adulteration, microbial contamination and the influence of feed on the quality of milk and dairy products. Because of the delays involved in the use of wet chemistry methods during the evaluation of these products, rapid analytical techniques such as near-infrared spectroscopy (NIRS) has gained prominence and proven to be an efficient tool, providing instant results. The technique is rapid, nondestructive, precise and cost-effective, compared with other laboratory techniques. Handheld NIRS devices are easily used on the farm to perform quality control measures on an incoming feed from suppliers, during feed preparation, milking and processing of cheese, butter and yoghurt. This ensures that quality feed, milk and other dairy products are obtained. This review considers research articles published in reputable journals which explored the possible application of NIRS in the dairy industry. Emphasis was on what quality parameters were easily measured with NIRS, and the limitations in some instances.

Novel Biomarkers of Mastitis in Goat Milk Revealed by MALDI-TOF-MS-Based Peptide Profiling

Mastitis is the most common infection of dairy goats impairing milk production and quality, which is usually recognized by mammary gland visual inspection and palpation. Subclinical forms of the disease are also widely represented, which lack the typical signs of the clinical ones but are still associated with reduced production and safety for human consumption of milk, generally presenting a high bacterial count. In order to obtain novel analytical tools for rapid and non-invasive diagnosis of mastitis in goats, we analyzed milk samples from healthy, subclinical and clinical mastitic animals with a MALDI-TOF-MS-based peptidomic platform, generating disease group-specific spectral profiles whose signal intensity and mass values were analyzed by statistics. Peculiar spectral signatures of mastitis with respect to the control were identified, while no significant spectral differences were observed between clinical and subclinical milk samples. Discriminant signals were assigned to specific peptides through nanoLC-ESI-Q-Orbitrap-MS/MS experiments. Some of these molecules were predicted to have an antimicrobial activity based on their strong similarity with homolog bioactive compounds from other mammals. Through the definition of a panel of peptide biomarkers, this study provides a very rapid and low-cost method to routinely detect mastitic milk samples even though no evident clinical signs in the mammary gland are observed.

Total conjugated linoleic acid content of ruminant milk: The world status insights

Conjugated linoleic acid (CLA) content of ruminant milk reported in published research papers (n = 65) from January 1995 to March 2020 around the world were analyzed to estimate the overall mean CLA value. The CLA content of ruminant milk samples was grouped according to geographical regions (Europe, South America, North America, Oceania, Asia, and Africa). The total CLA content of milk samples from cows, sheep, goats, yaks, and llama retrieved from the collected data ranged between 0.06 and 2.96% of total fatty acids. There is a wide variation of pooled estimated mean content of CLA in milk among the study regions and were highest in Oceania with 1.33% (95% confidence interval (CI): 1.16 - 1.49%) of total fatty acids. Though several factors have been reported to influence the CLA content of milk, the effect of the "geographical origin" was only considered in the present manuscript as one of the main factors in this respect.

Application of the Non-Destructive NIR Technique for the Evaluation of Strawberry Fruits Quality Parameters

The determination of strawberry fruit quality through the traditional destructive lab techniques has some limitations related to the amplitude of the samples, the timing and the applicability along all phases of the supply chain. The aim of this study was to determine the main qualitative characteristics through traditional lab destructive techniques and Near Infrared Spectroscopy (NIR) in fruits of five strawberry genotypes. Principal Component Analysis (PCA) was applied to search for spectral differences among all the collected samples. A Partial Least Squares regression (PLS) technique was computed in order to predict the quality parameters of interest. The PLS model for the soluble solids content prediction was the best performing-in fact, it is a robust and reliable model and the validation values suggested possibilities for its use in quality applications. A suitable PLS model is also obtained for the firmness prediction-the validation values tend to worsen slightly but can still be accepted in screening applications. NIR spectroscopy represents an important alternative to destructive techniques, using the infrared region of the electromagnetic spectrum to investigate in a non-destructive way the chemical-physical properties of the samples, finding remarkable applications in the agro-food market.

Analysis of the Acid Detergent Fibre Content in Turnip Greens and Turnip Tops (Brassica rapa L. Subsp. rapa) by Means of Near-Infrared Reflectance

Standard wet chemistry analytical techniques currently used to determine plant fibre constituents are costly, time-consuming and destructive. In this paper the potential of near-infrared reflectance spectroscopy (NIRS) to analyse the contents of acid detergent fibre (ADF) in turnip greens and turnip tops has been assessed. Three calibration equations were developed: in the equation without mathematical treatment the coefficient of determination (R²) was 0.91, in the first-derivative treatment equation R² = 0.95 and in the second-derivative treatment R² = 0.96. The estimation accuracy was based on RPD (the ratio between the standard deviation and the standard error of validation) and RER (the ratio between the range of ADF of the validation as a whole and the standard error of prediction) of the external validation. RPD and RER values were of 2.75 and 9.00 for the treatment without derivative, 3.41 and 11.79 with first-derivative, and 3.10 and 11.03 with second-derivative. With the acid detergent residue spectrum the wavelengths were identified and associated with the ADF contained in the sample. The results showed a great potential of NIRS for predicting ADF content in turnip greens and turnip tops.

FR4-based electromagnetic scanning micro-grating integrated with an angle sensor for a low-cost NIR micro-spectrometer

Aiming to implement a low-cost single-photodetector-based NIR micro-spectrometer, we present a novel flame retardant 4 (FR4)-driven micro-grating for spectral dispersion and spatial scanning. It consists of a silicon blazed grating directly bonded onto the FR4 actuator platform and an integrated differential electromagnetic angle sensor. Owing to the better shock and vibration reliability, larger aperture, shorter fabrication cycle, and much lower cost, it may be a potential alternative to conventional silicon microelectromechanical systems scanning micro-gratings. The micro-spectrometer prototype based on this device shows a spectral range of 800-1800 nm, a spectral accuracy of ±1.3  nm, and a 10 nm spectral resolution.

Applications of Photonics in Agriculture Sector: A Review

The agricultural industry has made a tremendous contribution to the foundations of civilization. Basic essentials such as food, beverages, clothes and domestic materials are enriched by the agricultural industry. However, the traditional method in agriculture cultivation is labor-intensive and inadequate to meet the accelerating nature of human demands. This scenario raises the need to explore state-of-the-art crop cultivation and harvesting technologies. In this regard, optics and photonics technologies have proven to be effective solutions. This paper aims to present a comprehensive review of three photonic techniques, namely imaging, spectroscopy and spectral imaging, in a comparative manner for agriculture applications. Essentially, the spectral imaging technique is a robust solution which combines the benefits of both imaging and spectroscopy but faces the risk of underutilization. This review also comprehends the practicality of all three techniques by presenting existing examples in agricultural applications. Furthermore, the potential of these techniques is reviewed and critiqued by looking into agricultural activities involving palm oil, rubber, and agro-food crops. All the possible issues and challenges in implementing the photonic techniques in agriculture are given prominence with a few selective recommendations. The highlighted insights in this review will hopefully lead to an increased effort in the development of photonics applications for the future agricultural industry.

Total and Free Fatty Acids Analysis in Milk and Dairy Fat

Dairy fat is one of the most complex natural fats because of its fatty acid (FA) composition. Ruminant dairy fat contains more than 400 different FA varying in carbon chain length, and degree, position and configuration of unsaturation. The following article reviews the different methods available to analyze FA (both total and free) in milk and dairy products. The most widely used methodology for separating and analyzing dairy FA is gas chromatography, coupled to a flame ionization detector (CG-FID). Alternatively, gas chromatography coupled to a mass spectrometer (GC-MS) is also used. After lipid extraction, total FA (TFA) are commonly converted into their methyl esters (fatty acid methyl esters, FAME) prior to chromatographic analysis. In contrast, free FA (FFA) can be analyzed after conversion to FAME or directly as FFA after extraction from the product. One of the key questions when analyzing FAME from TFA is the selection of a proper column for separating them, which depends mainly on the objective of the analysis. Quantification is best achieved by the internal standard method. Recently, near-infrared spectroscopy (NIRS), Raman spectroscopy (RS) and nuclear magnetic resonance (NMR) have been reported as promising techniques to analyze FA in milk and dairy products.

Evaluation of Yogurt Quality during Storage by Fluorescence Spectroscopy

The physico-chemical parameters including pH and viscosity, and the fluorescence signal induced by fluorescent compounds presenting in yogurts such as riboflavin and porphyrin were measured during one week’s storage at room temperature when five brands of yogurt samples were exposed to ambient air. The fluorescence spectra of yogurt showed four evident emission peaks, 525 nm, 633 nm, 661 nm, and 672 nm. To quantitatively investigate the quality of yogurt during deteriorating, a calculating method of the average rate of change (ARC) was proposed to study the relative change of fluorescence intensity in the spectral range of 600 to 750 nm associated with porphyrin and chlorin compounds. During the storage, the time evolution of two ARC, pH value, and viscosity were regular. Moreover, the ARC showed a good linear relationship with pH value and viscosity of yogurt. Further, multiple linear regression (MLR) models using two ARC as independent variables were developed to verify the dependence of fluorescence signal with pH value and viscosity, which showed a good linear relationship with an R-square of more than 85% for each class of yogurt. The results demonstrate that fluorescence spectra have a great potential to predict the quality of yogurt.

Application of visible and near-infrared spectroscopy for evaluation of ewes milk with different feeds

Rapid measurement optical techniques have the advantage over traditional methods of being faster and non-destructive. In this work visible and near-infrared spectroscopy (vis-NIRS) was used to investigate differences between measured values of key milk properties (e.g. fat, protein and lactose) in 30 samples of ewes milk according to three feed systems; faba beans, field peas and control diet. A mobile fibre-optic vis-NIR spectrophotometer (350–2500 nm) was used to collect reflectance spectra from milk samples. Principal component analysis was used to explore differences between milk samples according to the feed supplied, and a partial least-squares regression and random forest regression were adopted to develop calibration models for the prediction of milk properties. Results of the principal component analysis showed clear separation between the three groups of milk samples according to the diet of the ewes throughout the lactation period. Milk fat, protein and lactose were predicted with good accuracy by means of partial least-squares regression (R2 = 0.70–0.83 and ratio of prediction deviation, which is the ratio of standard deviation to root mean square error of prediction = 1.85–2.44). However, the best prediction results were obtained with random forest regression models (R2 = 0.86–0.90; ratio of prediction deviation = 2.73–3.26). The adoption of the vis-NIRS coupled with multivariate modelling tools can be recommended for exploring to differences between milk samples according to different feed systems, and to predict key milk properties, based particularly on the random forest regression modelling technique.

Comparison of Benchtop Fourier-Transform (FT) and Portable Grating Scanning Spectrometers for Determination of Total Soluble Solid Contents in Single Grape Berry (Vitis vinifera L.) and Calibration Transfer

Near-infrared (NIR) spectroscopy was applied for the determination of total soluble solid contents (SSC) of single Ruby Seedless grape berries using both benchtop Fourier transform (VECTOR 22/N) and portable grating scanning (SupNIR-1500) spectrometers in this study. The results showed that the best SSC prediction was obtained by VECTOR 22/N in the range of 12,000 to 4000 cm⁻¹ (833–2500 nm) for Ruby Seedless with determination coefficient of prediction (R_p²) of 0.918, root mean squares error of prediction (RMSEP) of 0.758% based on least squares support vector machine (LS-SVM). Calibration transfer was conducted on the same spectral range of two instruments (1000–1800 nm) based on the LS-SVM model. By conducting Kennard-Stone (KS) to divide sample sets, selecting the optimal number of standardization samples and applying Passing-Bablok regression to choose the optimal instrument as the master instrument, a modified calibration transfer method between two spectrometers was developed. When 45 samples were selected for the standardization set, the linear interpolation-piecewise direct standardization (linear interpolation-PDS) performed well for calibration transfer with R_p² of 0.857 and RMSEP of 1.099% in the spectral region of 1000–1800 nm. And it was proved that re-calculating the standardization samples into master model could improve the performance of calibration transfer in this study. This work indicated that NIR could be used as a rapid and non-destructive method for SSC prediction, and provided a feasibility to solve the transfer difficulty between totally different NIR spectrometers.