Portable Near-infrared Spectrometer Research Articles

Seasonal variation of antioxidant bioactive compounds in southern highbush blueberry leaves and non-destructive quality prediction in situ by a portable near-infrared spectrometer

Blueberry leaves (BBL) are a natural source with strong antioxidant activity, but bioactive compounds and their seasonal variation remain vague. Here, two major classes of compounds including four caffeoylquinic acids and eight flavonoids were identified in two southern highbush cultivars (“Lanmei” #1 and “Jewel”) grown in China. Major bioactive compounds were discovered using an online HPLC post-column derivatization system and determined as neochlorogenic acid (NeoCA), chlorogenic acid (CA), rutin, hyperoside, and isoquercitrin. CA contributed the most to the BBL antioxidant activity. “Lanmei” showed significant advantages in terms of rutin content and antioxidant activity over “Jewel” (P < 0.05). The highest CA content (CAC) of juvenile “Jewel” leaves reached 17.9%. July was the optimum harvest time for both cultivars after fruiting stage. Total phenolic content (TPC) and Trolox equivalent antioxidant capacity (TEAC) of fresh BBL were accurately predicted by a portable near-infrared (NIR) device in a rapid, low-cost, and non-destructive way in situ.

Rapid assessment of vanilla (Vanilla planifolia) quality parameters using portable near-infrared spectroscopy combined with random forest

The determination of moisture and vanillin content significantly influences the quality of vanilla. Currently, conventional chemical methods employed for assessing these parameters are time-consuming, involve complex sample preparation, are expensive, and environmentally unfriendly due to the use of chemical solutions. Portable Near-Infrared (NIR) spectroscopy emerges as a promising alternative, characterized by smaller dimensions and lower costs. This study investigates the performance of two portable NIR spectrometers with distinct wavelengths at 740–1070 nm and 1350–2550 nm, in conjunction with Random Forest (RF) and Partial Least Square (PLS) regression, and preprocessing techniques including min-max normalization, 1st derivative, standard normal variate (SNV), multiplicative scatter correction (MSC), 1st derivative + SNV, and 1st derivative + MSC for predicting moisture and vanillin content. At the wavelength range of 1350–2550 nm, RF coupled with 1st derivative produced the best moisture content prediction model with an R2 of 0.971, and RF paired with 1st derivative+SNV yielded the best vanillin content prediction with an R2 of 0.983. This work highlights that the integration of portable NIR and RF allows for rapid and non-destructive detection of moisture and vanillin content. This methodology provides a novel regression method for predicting vanilla qualities.

Near-Infrared Spectroscopy Combined with Fuzzy Improved Direct Linear Discriminant Analysis for Nondestructive Discrimination of Chrysanthemum Tea Varieties.

The quality of chrysanthemum tea has a great connection with its variety. Different types of chrysanthemum tea have very different efficacies and functions. Moreover, the discrimination of chrysanthemum tea varieties is a significant issue in the tea industry. Therefore, to correctly and non-destructively categorize chrysanthemum tea samples, this study attempted to design a novel feature extraction method based on the fuzzy set theory and improved direct linear discriminant analysis (IDLDA), called fuzzy IDLDA (FIDLDA), for extracting the discriminant features from the near-infrared (NIR) spectral data of chrysanthemum tea. To start with, a portable NIR spectrometer was used to collect NIR data for five varieties of chrysanthemum tea, totaling 400 samples. Secondly, the raw NIR spectra were processed by four different pretreatment methods to reduce noise and redundant data. Thirdly, NIR data dimensionality reduction was performed by principal component analysis (PCA). Fourthly, feature extraction from the NIR spectra was performed by linear discriminant analysis (LDA), IDLDA, and FIDLDA. Finally, the K-nearest neighbor (KNN) algorithm was applied to evaluate the classification accuracy of the discrimination system. The experimental results show that the discrimination accuracies of LDA, IDLDA, and FIDLDA could reach 87.2%, 94.4%, and 99.2%, respectively. Therefore, the combination of near-infrared spectroscopy and FIDLDA has great application potential and prospects in the field of nondestructive discrimination of chrysanthemum tea varieties.

Portable near infrared (NIR) spectrometer coupled with machine learning to classify milk with subclinical mastitis

Milk is one of the traditional foods in the human diet. However, subclinical mastitis infection in cows could compromise the nutritional composition of milk as well as the consumer safety. In this study, we investigated the possibility of implementing near infrared spectroscopy (NIRS), using a portable spectrometer, as a screening method to detect milk contaminated with subclinical mastitis. The spectra was acquired from a highly complex milk sample set (two seasons: winter and summer; two years: 2021 and 2022; from five geographical locations within Brazil: Goiás, Pará, Paraíba, São Paulo, and Santa Catarina; different feeding systems.) were analyzed using Principal Component Analysis (PCA) and Partial Least Squares Discriminant Analysis (PLS-DA), as well as machine learning models like Random Forest (RF) and Support Vector Machine (SVM). PCA revealed that the clustering of mastitis milk and non-mastitis milk samples was associated with variations in lactose content. This observation was also noted in PLS-DA, which achieved accuracy values of 78%, compared to the 62% achieved by RF and SVM in detecting mastitis milk. However, the sensitivity (recall) for detecting mastitis samples was higher for RF (78%), and SVM notably excelled in detecting non-mastitis milk (81%). Furthermore, it was demonstrated that the removal of outliers using Isolation Forest enhances the performance of RF and SVM models based on NIR spectra, with increases of up to 25% in the precision. In conclusion, the portable NIR spectrometer could potentially serve as a screening method in the dairy industry to detect mastitis milk samples.

Integrating portable NIR spectrometry with deep learning for accurate Estimation of crude protein in corn feed

This study investigates the challenges encountered in utilizing portable near-infrared (NIR) spectrometers in agriculture, specifically in developing predictive models with high accuracy and robust generalization abilities despite limited spectral resolution and small sample sizes. The research concentrates on the near-infrared spectra of corn feed, utilizing spectral processing techniques and CNNs to precisely estimate crude protein content. Five preprocessing methods were implemented alongside two-dimensional (2D) correlation spectroscopy, resulting in the development of both one-dimensional (1D) and 2D regression models.A comparative analysis of these models in predicting crude protein content demonstrated that 1D-CNNs exhibited superior predictive performance within the 1D category. For the 2D models, CropNet and CropResNet were utilized, with CropResNet demonstrating more accurate and superior predictive capabilities. Overall, the integration of 2D correlation spectroscopy with suitable preprocessing techniques in deep learning models, particularly the 2D CropResNet, proved to be more precise in predicting the crude protein content in corn feed. This finding emphasis the potential of this approach in the portable spectrometer market.

A Multi-Scale Convolutional Neural Network Combined with a Portable Near-Infrared Spectrometer for the Rapid, Non-Destructive Identification of Wood Species

The swift and non-destructive classification of wood species holds crucial significance for the utilization and trade of wood resources. Portable near-infrared (NIR) spectrometers have the potential for rapid and non-destructive wood species identification, and while several studies have explored related methodologies, further research on their practical application is needed. To address this research gap, this study proposes a multi-scale convolutional neural network (CNN) combined with a portable NIR spectrometer (wavelengths range: 908 to 1676 nm) for wood species identification. To enhance the capability of directly extracting robust features from NIR spectral data collected by a portable spectrometer, the Gramian angular field (GAF) method is introduced to transform 1-dimensional (1D) NIR spectral data into 2-dimensional (2D) data matrices. Furthermore, a multi-scale CNN model is utilized for direct feature extraction. The representation by 2D matrices, instead of 1D NIR spectral data, aligns with 2D convolutional operations and enables a more robust extraction of discriminative features. In the experimental phase, eight wood species were identified using the proposed method, alongside commonly used multivariate data analysis and machine learning (ML) methods. The StratifiedGroupKFold dataset partitioning approach and five-fold cross-validation were used. Additionally, nine spectral preprocessing methods were compared, and principal component analysis (PCA) was used for feature extraction in the ML method. Evaluation metrics, such as accuracy, precision, and recall, were adopted to assess the performance of the methods. The proposed multi-scale CNN model, in combination with 2D GAF matrices of the 1D spectral data, yielded the most accurate results with a mean accuracy of 97.34% in the five-fold validation. These findings present a new approach for the construction of a rapid, non-destructive, and automatic wood species identification method using a portable NIR spectrometer.

Geographical origin discrimination of black beans using near-infrared (NIR) spectroscopy combined with fuzzy regularized complete linear discriminant analysis (FRCLDA)

Black beans from different geographical origins exhibit variations in terms of their nutritional and economic values. Therefore, ensuring traceability of the geographical origin of black beans is essential for both consumers and the product processing industries. The study designed a comprehensive black bean origin identification model using near-infrared (NIR) spectroscopy combined with discriminant analysis algorithms. To enhance the performance of the model, this study incorporated fuzzy logic into regularized complete linear discriminant analysis (RCLDA) and proposed the fuzzy regularized complete linear discriminant analysis (FRCLDA). This innovative approach aims to extract additional information from overlapping NIR spectra of black beans. NIR spectra of black beans from five locations were collected employing a portable near-infrared spectrometer. In the preprocessing stage, the Savitzky–Golay (SG) and mean centering (MC) were applied to process the raw spectra. Subsequently, complete linear discriminant analysis (CLDA), RCLDA, and FRCLDA were utilized to extract the information from the processed data. Finally, k-nearest neighbor (KNN), support vector machines (SVM), and extreme learning machine (ELM) were employed to classify the black bean samples based on the extracted features. The results indicated that KNN achieved the best classification performance. The FRCLDA-KNN model achieved the exceptional performance with a classification accuracy of 99.23% on the test set. Therefore, the system demonstrates satisfactory accuracy and performance in accurately identifying black beans from different geographical origins.

Qualitative discrimination and quantitative prediction of microplastics in ash based on near-infrared spectroscopy

Microplastics are recognized as a new environmental pollutant. Researchers have detected their presence in waste incineration ash. However, traditional testing methods take a very long testing period. There is a lack of research on detecting microplastics in waste incineration ash. In this paper, a portable near-infrared spectra (NIRS) spectrometer was used for qualitative discrimination and quantitative prediction of microplastics in ash. A total of 84 sets of simulated ash samples containing different types (PP, PS, PE, and PVC) and contents (2.4 wt% − 20 wt%) of microplastics were used in the model. The results show the qualitative discrimination model using support vector machines (SVM) method with multiplicative scatter correction (MSC) preprocessing could effectively identify the microplastic types in the ash with 100% detection accuracy. Furthermore, the partial least squares regression (PLSR) model was effective in quantitatively predicting the content of microplastics in ash. The Rp2 of the PP, PS, PE, and PVC models are 0.95, 0.93, 0.89, and 0.95, respectively. The RPD of the PP, PS, PE, and PVC models are 3.97, 3.96, 2.89 and 5.02, respectively. This study shows that microplastics in ash can be detected rapidly and accurately using portable near-infrared spectrometers.

In situ rapid evaluation method of quality of peach kernels based on near infrared spectroscopy

This study aimed to perform a rapid in situ assessment of the quality of peach kernels using near infrared (NIR) spectroscopy, which included identifications of authenticity, species, and origins, and amygdalin quantitation. The in situ samples without any pretreatment were scanned by a portable MicroNIR spectrometer, while their powder samples were scanned by a benchtop Fourier transform NIR (FT-NIR) spectrometer. To improve the performance of the in situ determination model of the portable NIR spectrometer, the two spectrometers were first compared in identification and content models of peach kernels for both in situ and powder samples. Then, the in situ sample spectra were transferred by using the improved principal component analysis (IPCA) method to enhance the performance of the in situ model. After model transfer, the prediction performance of the in situ sample model was significantly improved, as shown by the correlation coefficient in the prediction set (Rp), root means square error of prediction (RMSEP), and residual prediction deviation (RPD) of the in situ model reached 0.9533, 0.0911, and 3.23, respectively, and correlation coefficient in the test set (Rt) and root means square error of test (RMSET) reached 0.9701 and 0.1619, respectively, suggesting that model transfer could be a viable solution to improve the model performance of portable spectrometers.

In-situ detection on near-infrared spectra fingerprints of asphalt mixture after laboratory short- and long-term aging

Although Fourier transform infrared spectroscopy has been widely used in research on the aging of asphalt binders, it is a cumbersome, time-consuming, and environmentally polluting process using toxic solvents to extract and recover asphalt binders from asphalt mixtures for spectral measurement and analysis. However, little effort has been devoted towards its non-destructive spectral measurement, which is important for intelligent monitoring of pavement deterioration conditions. This paper focused on the non-destructive investigation of comparative near-infrared spectra of asphalt mixture specimens designed using the Marshall method after laboratory short- and long-term aging with a portable near-infrared spectrometer. The near-infrared spectra showed that asphalt mixtures from different aging states appeared significantly different fingerprints. The near-infrared fingerprints and aging states were also statistically significantly different according to the analysis of one-way analysis of variance. These distinctive near-infrared fingerprints could be attributed to the varied amounts of chemical functional groups, mainly hydrogen-containing groups during oxidative aging. This study provided a non-destructive and in-situ detection of the aging of asphalt mixtures.

Near-infrared spectroscopy as a novel method of exvivo bladder cancer tissue characterisation.

To evaluate near-infrared (NIR) spectroscopy in differentiating between benign and malignant bladder pathologies ex vivo immediately after resection, including the grade and stage of malignancy. A total of 355 spectra were measured on 71 bladder specimens from patients undergoing transurethral resection of bladder tumour (TURBT) between April and August 2022. Scan time was 5 s, undertaken using a portable NIR spectrometer within 10 min from excision. Specimens were then sent for routine histopathological correlation. Machine learning models were applied to the spectral dataset to construct diagnostic algorithms; these were then tested for their ability to predict the histological diagnosis of each sample using its NIR spectrum. A two-group algorithm comparing low- vs high-grade urothelial cancer demonstrated 97% sensitivity, 99% specificity, and the area under the receiver operating characteristic curve (AUC) was 0.997. A three-group algorithm predicting stages Ta vs T1 vs T2 achieved 97% sensitivity, 92% specificity, and the AUC was 0.996. This first study evaluating the diagnostic potential of NIR spectroscopy in urothelial cancer shows that it can be accurately used to assess tissue in an ex vivo setting immediately after TURBT. This offers point-of-care assessment of bladder pathology, with potential to influence the extent of resection, reducing both the need for re-resection where invasive disease may be suspected, and also the potential for complications where extent of diagnostic resection can be limited. Further studies utilising fibre-optic probes offer the potential for in vivo assessment.

Rapid and non-destructive quality estimation of cinchona, andrographis paniculata, and black pepper using a portable NIR spectroscopy measuring device

Quality assessment of natural products are mostly done by laboratory methods which are invasive, cost-intensive, with low-throughput and prone to human error. Near infrared (NIR) spectroscopy have emerged as a very useful technique as a rapid non-destructive detection and quality estimation tool in identifying chemical fingerprint sensitive to the vibrational modes in chemical bonds like O-H, C-H and N-H.In this work, a reasonably low-cost and portable NIR spectrometer was developed with a few imported internal components, together with compatible software for its use, like the graphical user interface (GUI) and calibration program. Quinine in Cinchona (Cinchona L) bark, andrographolides in Andrographis paniculata and piperine in black pepper were estimated and the coefficient of determination (R2) of 0.93, 0.94, and 0.90 was obtained respectively for the regression analysis. The model for cinchona, andrographolides, and piperine gave residual prediction deviation (RPDs) of 4.25, 4.21 and 3.0, range error ratio (RERs) of 14.59, 14.42, 10.32 and the ratio between the standard error of calibration (SEC) and the standard error of prediction (SEP) i.e. SEP/SEC of 1.18, 0.21 and 0.92 respectively. The derived partial least squares (PLS) regression models fulfil the requirements of AACC Method 39–00 (AACC in AACC Method, 39–00:15, 1999) used for screening (RPD ≥ 2.5), as per knowledge of the authors. The device can be used for other molecular markers in plants and their byproducts.

Detecting Respiratory Viruses Using a Portable NIR Spectrometer-A Preliminary Exploration with a Data Driven Approach.

Respiratory viruses' detection is vitally important in coping with pandemics such as COVID-19. Conventional methods typically require laboratory-based, high-cost equipment. An emerging alternative method is Near-Infrared (NIR) spectroscopy, especially a portable one of the type that has the benefits of low cost, portability, rapidity, ease of use, and mass deployability in both clinical and field settings. One obstacle to its effective application lies in its common limitations, which include relatively low specificity and general quality. Characteristically, the spectra curves show an interweaving feature for the virus-present and virus-absent samples. This then provokes the idea of using machine learning methods to overcome the difficulty. While a subsequent obstacle coincides with the fact that a direct deployment of the machine learning approaches leads to inadequate accuracy of the modelling results. This paper presents a data-driven study on the detection of two common respiratory viruses, the respiratory syncytial virus (RSV) and the Sendai virus (SEV), using a portable NIR spectrometer supported by a machine learning solution enhanced by an algorithm of variable selection via the Variable Importance in Projection (VIP) scores and its Quantile value, along with variable truncation processing, to overcome the obstacles to a certain extent. We conducted extensive experiments with the aid of the specifically developed algorithm of variable selection, using a total of four datasets, achieving classification accuracy of: (1) 0.88, 0.94, and 0.93 for RSV, SEV, and RSV + SEV, respectively, averaged over multiple runs, for the neural network modelling of taking in turn 3 sessions of data for training and the remaining one session of an 'unknown' dataset for testing. (2) the average accuracy of 0.94 (RSV), 0.97 (SEV), and 0.97 (RSV + SEV) for model validation and 0.90 (RSV), 0.93 (SEV), and 0.91 (RSV + SEV) for model testing, using two of the datasets for model training, one for model validation and the other for model testing. These results demonstrate the feasibility of using portable NIR spectroscopy coupled with machine learning to detect respiratory viruses with good accuracy, and the approach could be a viable solution for population screening.

Nondestructive prediction of lime acidity with a single scan using two types of near infrared spectrometers and ensemble learning strategy

An accurate and nondestructive prediction method of lime acidity was established based on near infrared (NIR) spectroscopy and ensemble learning strategy. Dual-band spectra were obtained nondestructively with a single scan using a desk-top Fourier transform spectrometer and a grating portable spectrometer. Spectral preprocessing methods were used to eliminate the interferences in the spectra. The quantification models of available acidity (pH and 10−pH) and total acidity (TA) were developed with the ensemble learning strategy compared with partial least squares (PLS) and variable selection methods. The results indicated that due to the high-energy light source, the models of the grating portable spectrometer were much better than those of the Fourier transform spectrometer. Short-wave NIR (SWNIR) was more suitable for quantitative analysis of available acidity, while long-wave NIR (LWNIR) was more effective for quantitative analysis of TA. Besides, the models of available acidity were ahead of those of TA. Compared with PLS and variables selection methods, the ensemble learning strategy can produce models with higher prediction accuracy and better robustness. In the optimized models, the correlation coefficients of pH, 10−pH and TA for the prediction set were 0.84, 0.82 and 0.66, respectively. The experiment results show that accurate and nondestructive prediction of lime acidity can be achieved with the grating portable NIR spectrometer and ensemble learning strategy.

A comprehensive “quality-quantity-activity” approach based on portable near-infrared spectrometer and membership function analysis to systematically evaluate spice quality: Cinnamomum cassia as an example

Spices have long been popular worldwide. Besides serving as aromatic and flavorful food and cooking ingredients, many spices exhibit notable bioactivity. Quality evaluation methods are essential for ensuring the quality and flavor of spices. However, existing methods typically focus on the content of particular components or certain aspects of bioactivity. For a systematic evaluation of spice quality, we herein propose a comprehensive “quality-quantity-activity” approach based on portable near-infrared spectrometer and membership function analysis. Cinnamomum cassia was used as a representative example to illustrate this approach. Near-infrared spectroscopy and chemometric methods were combined to predict the geographical origin, cinnamaldehyde content, ash content, antioxidant activity, and integrated membership function value. All the optimal prediction models displayed good predictive ability (correlation coefficient of prediction > 0.9, residual predictive deviation > 2.1). The proposed approach can provide a valuable reference for the rapid and comprehensive quality evaluation of spices.

Non-destructive Identification of the geographical origin of red jujube by near-infrared spectroscopy and fuzzy clustering methods

ABSTRACT The red jujube quality is closely associated with its place of origin. In order to quickly and easily identify the geographical origin of red jujube, the classification of red jujube samples’ near-infrared reflectance (NIR) spectra was performed using several fuzzy clustering methods in combination with principal component analysis (PCA) and linear discriminant analysis (LDA). Firstly, a NIR-M-R2 portable near-infrared spectrometer was used to collect four varieties of red jujube samples from four representative producing areas in four provinces: Gansu, Henan, Shanxi and Xinjiang in China. Each variety corresponded to a producing area, and it had 60 samples with a total of 240 samples. Near-infrared spectra of red jujube were acquired using a NIR-M-R2 portable near-infrared spectrometer, and the initial near-infrared spectra were preprocessed by Savitzky-Golay (SG) filtering. Secondly, PCA and LDA were used to further process the NIR data for dimension reduction and feature extraction, respectively. Finally, red jujube samples were classified by fuzzy C-means (FCM) clustering, Gustafson-Kessel (GK) clustering and possibility fuzzy C-means (PFCM) clustering. When GK served as the clustering algorithm, the clustering accuracy was the highest, as the value of 98.8%. Based on the experimental results, it was evident that the GK clustering algorithm played a significant role in identifying the place of origin of red jujube with near-infrared spectroscopy.

Quantification of caffeine and catechins and evaluation of bitterness and astringency of Pu-erh ripen tea based on portable near-infrared spectroscopy

Rapid and accurate assaying catechins, caffeine, bitterness and astringency prediction are integral to quality assessment and control of Pu-erh ripen tea (PRT). The traditional experimental procedures for content quantification and sensory evaluation are time-consuming and lab-dependent, while near-infrared (NIR) technology has the potential to address this issue. In this work, an efficient and environment-friendly approach was established for caffeine and catechins prediction, and bitterness and astringency evaluation of PRT, using a portable NIR spectrometer coupled with chemometrics. A total of 100 PRT samples were collected for analysis and seven different spectral preprocessing methods and two variable selection algorithms were employed to improve the prediction performance of the partial least squares regression (PLSR) model. The optimized models achieved satisfactory results for the prediction of caffeine and catechins (catechin, catechin gallate, gallocatechin, epicatechin gallate, epigallocatechin, and epigallocatechin gallate), as well as bitterness and astringency, with the correlation coefficient of the prediction set (Rp) above 0.9, and residual prediction deviation (RPD) over 2.5. This research offers an alternative for portable and rapid quantification of caffeine and eight catechins and evaluation of bitterness and astringency of PRT.

Application of Near-Infrared Spectroscopy and Fuzzy Improved Null Linear Discriminant Analysis for Rapid Discrimination of Milk Brands.

The quality of milk is tightly linked to its brand. A famous brand of milk always has good quality. Therefore, this study seeks to design a new fuzzy feature extraction method, called fuzzy improved null linear discriminant analysis (FiNLDA), to cluster the spectra of collected milk for identifying milk brands. To elevate the classification accuracy, FiNLDA was applied to process the near-infrared (NIR) spectra of milk acquired by the portable near-infrared spectrometer. The principal component analysis and Savitzky-Golay (SG) filtering algorithm were employed to lower dimensionality and eliminate noise in this system, respectively. Thereafter, improved null linear discriminant analysis (iNLDA) and FiNLDA were applied to attain the discriminant information of the NIR spectra. At last, the K-nearest neighbor classifier was utilized for assessing the performance of the identification system. The results indicated that the maximum classification accuracies of LDA, iNLDA and FiNLDA were 74.7%, 88% and 94.67%, respectively. Accordingly, the portable NIR spectrometer in combination with FiNLDA can classify milk brands correctly and effectively.

Rapid assessment of traditional Chinese condiment of chicken essence by portable NIRS and Raman modeling

As a commonly used complex condiment in daily cooking, the chicken essence occupies a large proportion of the condiment market in China. However, in recent years, quality issues with condiments, including chicken essence, have negatively impacted the economic interests of both the market and consumers. Thus, in this study, two types of portable near infrared spectrometer (NIRS) and one portable Raman spectrometer were used to establish the quantitative model of disodium 5′-ribonucleotide (IMP+GMP) in chicken essence powder. Results showed that the three portable spectrometers can successfully realize the quantification of the key component in chicken essence; the Rc2 and Rp2 of the established PLS model of handheld NIR spectroscopy was 0.9527 and 0.9455, respectively. The RPD value reached 4.2328, indicating the model exhibits a satisfactory predictability. The optimized Raman model also performed well with an RPD value of 3.6571. The successful rapid detection of the quality component of chicken essence can further explore the potential of miniature molecular spectroscopy in non-destructive detection of other complex powdered condiments.

Achieving prediction of starch in cassava (Manihot esculenta Crantz) by data fusion of Vis-NIR and Mid-NIR spectroscopy via machine learning

Cassava is an important crop for both the food and bio-energy industries, serving as a primary carbohydrate source and a staple food. Knowledge of the starch content (SC) is a key parameter index its quality. Non-destructive measurement of SC is needed to track the same tubers to study behaviours of SC accumulation in cassava tubers, which benefit for breeders in discovering a good variety and farmer to indicating harvesting period. This paper involves the prediction of cassava tuber starch content (SC) using multiple near infrared (NIR) spectrometers, aiming to measure SC in fresh cassava tuber. This study applies both portable NIR spectrometers at 570–1031 nm and 860 – 1760 nm. The best results of the model which developed provided R2p and RPD were 0.69 and 1.80 for VIS-NIR region. Meanwhile, the Mid-NIR region provided poor performance with R2p and RPD of 0.46 and 1.36, respectively. Therefore, this study aims to assess whether it was better for predicting SC if the model was developed using combined selected significant wavelength from both VIS-NIR and Mid-NIR regions using machine learning. Genetics algorithm (GA) and variables important projection (VIP) were used for selecting significant wavelength. Selected wavelengths were combined and then generate calibration models via machine learning (ML). The results of the SC model developed using NIR spectra from selected wavelength with SV regression method had the highest performance, with R2c and RMSEc of 0.88 and 1.28%. Meanwhile, the R2p, RMSEp, and RPD of 0.74, 1.86% and 1.97, respectively. Then, the best calibration model was used to measure the unknown sample which corrected from a different harvest season. The external test set provided R2p, RMSEp were 0.88 and 1.38%. The results indicate that combination of NIR with different region can achieve.