RPD Values Research Articles

NIR quantitative model trans-scale calibration from small scale to pilot scale via directed DOSC-SBC algorithm

In order to solve the problem of inapplicability of NIR quantitative models due to the large difference between the modeling samples and the samples to be tested, Directed DOSC-SBC（DDOSC-SBC）algorithm is proposed in this paper based on Direct Orthogonal Signal Correction combined with Slope/Bias Correction (DOSC-SBC) algorithm. To obtain the suitable spectral matrix transfer parameters for the test set during DDOSC spectral preprocessing, several representative test samples in the test set were selected, then the spectral systematic errors between the modeling set and the test set were corrected with the SBC method in order to realize the trans-scale prediction of the NIR quantitative model. NIR data and the critical quality attributes（CQAs）were detected in the small scale and pilot scale pharmaceutical process of the fluidized bed granulation of dextrin and water extraction of honeysuckle. After the small scale model was calibrated via the directed DOSC-SBC algorithm which was guided by representative pilot scale samples, the small scale model was able to predict the pilot scale test samples more accurately. The NIR quantitative model trans-scale calibration from small scale to pilot scale was also successfully realized with a RPD value higher than 3.5 and RSEP value lower than 10%. DDOSC-SBC algorithm is a successful model trans-scale calibrated method that can be applied to NIR real-time monitoring of CQAs in the preparation process of Chinese herbal medicine.

Analysis of the spectral and textural features of hyperspectral images for the nondestructive prediction of amylopectin and amylose contents of sorghum

The contents of amylopectin and amylose are important indicators to evaluate the quality of sorghum. Hyperspectral imaging (HSI) is a technology based on machine vision and spectroscopy that has been widely used in the quality evaluation of various foods. Three combined algorithms were used to extract the characteristic wavelengths. The gray-level co-occurrence matrix (GLCM) was used to extract the textural features. Based on the spectra, textures, and fusion data, genetic algorithm-optimized back propagation neural network (BPNN-GA), cascade forest (CF), and partial least square regression (PLSR) models were established to predict the amylopectin and amylose contents. The BPNN-GA and CF models established using the characteristic wavelengths of visible-light were determined to be the best model, with RPD values of 8.2352 and 26.7308, respectively. These results demonstrated that HSI combined with machine learning algorithms could be used to enable the rapid and nondestructive prediction of amylopectin and amylose contents in sorghum.

Qualitative and Quantitative Detection of Acacia Honey Adulteration with Glucose Syrup Using Near-Infrared Spectroscopy

Honey adulteration with cheap sweeteners such as corn syrup or invert syrup results in honey of lesser quality that can harm the objectives of both manufacturers and consumers. Therefore, there is a growing interest for the development of a fast and simple method for adulteration detection. In this work, near-infrared spectroscopy (NIR) was used for the detection of honey adulteration and changes in the physical and chemical properties of the prepared adulterations. Fifteen (15) acacia honey samples were adulterated with glucose syrup in a range from 10% to 90%. Raw and pre-processed NIR spectra of pure honey samples and prepared adulterations were subjected to Principal Component Analysis (PCA), Partial Least Squares (PLS) regression, and Artificial Neural Network (ANN) modeling. The results showed that PCA ensures distinct grouping of samples in pure honey samples, honey adulterations, and pure adulteration using NIR spectra after the Multiplicative Scatter Correction (MSC) method. Furthermore, PLS models developed for the prediction of the added adulterant amount, moisture content, and conductivity can be considered sufficient for screening based on RPD and RER values (1.7401 < RPD < 2.7601; 7.7128 < RER < 8.7157) (RPD of 2.7601; RER of 8.7157) and can be moderately used in practice. The R2validation of the developed ANN models was greater than 0.86 for all outputs examined. Based on the obtained results, it can be concluded that NIR coupled with ANN modeling can be considered an efficient tool for honey adulteration quantification.

Fusion of three spectroscopic techniques for prediction of fatty acid in processed lamb

The application of individual spectroscopic techniques for meat analysis has been widely explored. Attempts to fuse data from multiple spectroscopic instruments for meat analysis are still lacking. Comparative assessment of the performance of mid infrared (MIR), near infrared (NIR) and Raman spectroscopy to estimate fatty acid (FA) composition in processed lamb was investigated. The acquired data from these individual techniques were then utilised in estimating similar parameters using a multi-block partial least square data fusion approach. Model performance was assessed with respect to the determination coefficient and ratio of predictive deviation upon cross-validation of the model. The fused data had slight improvements for the prediction of four FA parameters including MUFA, C18:0, C18:1 c9 and C9, t11- CLA), suggesting possible information enhancement with use of multiple instruments. However, MIR offered better predictability (RPD values) across the FA parameters considered.

Development and optimization of NIRS prediction models for simultaneous multi-trait assessment in diverse cowpea germplasm.

Cowpea (Vigna unguiculata (L.) Walp.) is one such legume that can facilitate achieving sustainable nutrition and climate change goals. Assessing nutritional traits conventionally can be laborious and time-consuming. NIRS is a technique used to rapidly determine biochemical parameters for large germplasm. NIRS prediction models were developed to assess protein, starch, TDF, phenols, and phytic acid based on MPLS regression. Higher RSQexternal values such as 0.903, 0.997, 0.901, 0.706, and 0.955 were obtained for protein, starch, TDF, phenols, and phytic acid respectively. Models for all the traits displayed RPD values of >2.5 except phenols and low SEP indicating the excellent prediction of models. For all the traits worked, p-value ≥ 0.05 implied the accuracy and reliability score >0.8 (except phenol) ensured the applicability of the models. These prediction models will facilitate high throughput screening of large cowpea germplasm in a non-destructive way and the selection of desirable chemotypes in any genetic background with huge application in cowpea crop improvement programs across the world.

A portable smart detection and electrocatalytic mechanism of mycophenolic acid: A machine learning-based electrochemical nanosensor to adapt variable-pH silage microenvironment

The pH of silage microenvironment is protean which affects the concentration of mycophenolic acid (MPA) that is an animal health-threatening mycotoxin produced by Penicillium roqueforti. This inspired us to develop a fast portable intelligent method for electrochemical detection of MPA in silage with the variable-pH microenvironment using Zn-Co MOF/Ti3C2 MXene/Fe3O4-MGO coupling with machine learning (ML). Zn-Co MOF (metal organic framework), Ti3C2 MXene (graphene-like titanium carbide MXene), Fe3O4-MGO (magnetic Fe3O4-graphene oxide) and their nanocompsite with excellent electrocatalytic capacity were prepared and characterized. The electrocatalytic mechanism of MPA was investigated by density functional theory (DFT) and electrochemical experiments, which clarified the most easily redox position of MPA. ML model via artificial neural network (ANN) algorithm for smart output of MPA through input of pH was discussed that adapt to the variable-pH microenvironment and realize intelligent analysis of MPA in silage with the variable-pH microenvironment. R2 near 1, lower both RMSE and MAE, and higher RPD value demonstrate the good predictive performance and high predictive accuracy of the proposed ANN model. This will provide a fast portable wireless intelligent sensing analytical technology for detecting hazardous substances in diverse complicated and changeable outdoor microenvironments

Nondestructive testing of the quality of different grades of creamy strawberries based on Laida algorithm

Rapid nondestructive testing of fruit quality is one of the hotspots in food industry research. This paper studied the quality detection of different grades of cream strawberries. Preprocessing methods such as detrending, moving average smoothing (MA), and standard normal variables (SNV) were used to eliminate spectral data errors. The competitive adaptive reweighted sampling algorithm (CARS), successive projection algorithm (SPA), the combination of the above two algorithms (CARS+SPA), and the self-programming algorithm based on the Laida criterion (collectively referred to as the Laida algorithm) were used to reduce the dimension of data, and a partial least squares regression prediction model was established. The results show that when the Laida algorithm predicted the soluble solid content (SSC), total acid (TA), and vitamin C (VC) content of strawberries, the prediction set correlation coefficients were 0.919, 0.931, and 0.907, respectively, the values of RPD were 3.15, 4.00, and 3.44, respectively, the relative errors of the predicted values and the measured values were 3.93%, 5.74%, and 3.69%, respectively. Practical applications This shows that it is feasible to use the Laida algorithm to extract the characteristic wavelengths and establish a prediction model for the SSC, TA, and VC content in cream strawberries. This study can provide a theoretical basis for the development of a rapid detection instrument for the quality of creamy strawberries.

Estimation of Soil Organic Carbon Content in Coastal Wetlands with Measured VIS-NIR Spectroscopy Using Optimized Support Vector Machines and Random Forests

Coastal wetland soil organic carbon (CW-SOC) is crucial for both “blue carbon” and carbon sequestration. It is of great significance to understand the content of soil organic carbon (SOC) in soil resource management. A total of 133 soil samples were evaluated using an indoor spectral curve and were categorized into silty soil and sandy soil. The prediction model of CW-SOC was established using optimized support vector machine regression (OSVR) and optimized random forest regression (ORFR). The Leave-One-Out Cross-Validation (LOO-CV) method was used to verify the model, and the performance of the two prediction models, as well as the models’ stability and uncertainty, was examined. The results show that (1) The SOC content of different coastal wetlands is significantly different, and the SOC content of silty soils is about 1.8 times that of sandy soils. Moreover, the characteristic wavelengths associated with SOC in silty soils are mainly concentrated in the spectral range of 500–1000 nm and 1900–2400 nm, while the spectral range of sandy soils is concentrated in the spectral range of 600–1400 nm and 1700–2400 nm. (2) The organic carbon prediction model of silty soil based on the OSVR method under the first-order differential of reflectance (R′) is the best, with the Adjusted-R2 value as high as 0.78, the RPD value is much greater than 2.0 and 5.07, and the RMSE value as low as 0.07. (3) The performance of the OSVR model is about 15~30% higher than that of the support vector machine regression (SVR) model, and the performance of the ORFR model is about 3~5% higher than that of the random forest regression (RFR) model. OSVR and ORFR are better methods of accurately predicting the CW-SOC content and provide data support for the carbon cycle, soil conservation, plant growth, and environmental protection of coastal wetlands.

Estimation of Moisture and Protein Content in Pumpkin Seeds Using NIRS with Partial Least Square (PLS) Method

Most people do not use pumpkin seeds and only fruit for food production. Meanwhile, pumpkin seeds contain antioxidants, protein, carbohydrates, and other vitamins. Measurement of the chemical content of a material is usually carried out destructively (conventional methods) in the laboratory. This method is expensive, takes a long time to repair samples, and creates chemical waste. One method that is currently being developed to detect the chemical content of a material is the near-infrared spectroscopy (NIRS) method. Near-infrared spectroscopy (NIRS) is a non-destructive method used to quickly analyze and obtain information on a material's chemical content without using chemicals. The NIRS data calibration technique using partial least squares (PLS) is a step carried out to build a model that relates the response spectra of each sample at each wavelength to chemical concentrations known from laboratory analysis and protein content in pumpkin seeds with NIRS using PLS calibration. Spectrum data pre-treatment was carried out with GapDerivarive and Derivative Savitzky-Golay. The results showed that the best calibration model for moisture and protein content was obtained using DerivativeGap data processing with values of r = 0.92, R2 = 0.84, SEC = 0.11%, and RMSEC = 0.11%. RPD = 2.66 and using Latent Variable (LV) factor 2, while for protein content r = 0.95, R2 = 0.89, SEC = 0.71%, and RMSEC = 0.70%. RPD value = 3.33 and using Latent Variable (LV) factor 3.

Pedological Development of Soils in Eastern Agroclimatic Zone of Haryana through Field Morphology Rating System

A study was carried out to evaluate the pedological development of soils in Eastern agroclimatic zone of Haryana using field morphology rating system. For this, nine pedons were excavated to study the morphological properties in field and rating scale was used to compare adjacent horizons with each other to give a comparison of the relative distinctness of horizons or to compare horizons in the solum with the C-horizon to give a relative profile development in soils occurring in different physiographic positions i.e. Shiwalik hills, recent and old alluvial plains with different precipitations throughout the study area. Some chemical properties were also used to evaluate the pedogenic factors as a result of being affected by process and factors of soil formation due to the distinct effect of climate. The Relative Horizons Distinctness (RHD) ratings were made by a comparison of adjacent horizons. The soils of Shiwalik hills, recent and old alluvial plains have RHD ratings from 6 to 14, 3 to 11 & 3 to 12, respectively. The RPD ratings were made by a comparison of the C horizon to the horizons above it in the profile. The soils of Shiwalik hills, recent and old alluvial plains have RPD ratings from 7 to 15, 3 to 14 & 5 to 16. The RPD values of all the profiles were maximum in A horizon due to maximum pedological development influenced by weathering. So, studies of these soils indicated the pedological development in the order of Shiwalik hills > old alluvial plains > recent alluvial plains.

Determination of protein content in single black fly soldier (Hermetia illucens L.) larvae by near infrared hyperspectral imaging (NIR-HSI) and chemometrics

The production of alternative proteins to meet the demand of a growing population has accelerated the growth of the market for edible insects. Black fly soldier (BFS) larvae (Hermetia illucens L.) have been widely studied globally due to their high content of fat, protein, and minerals, being mainly used for animal feed. Chemical analysis for determination of its composition is time consuming and laborious. In this work, we have developed predictive models based on Near Infrared Hyperspectral Imaging (NIR-HSI), Partial Least Square Regression (PLSR) and Support Vector Machine Regression (SVMR) to estimate the total protein content in single and intact BFS larvae. A variable selection step by interval PLS (iPLS) and genetic algorithms (GA) was implemented to improve regression model performance. In addition, BFS larvae hyperspectral images were explored using Principal Component Analysis (PCA), whose results showed the distribution of the different chemical compounds in the larvae. The PLSR and SVMR models reached RMSEP values of 1.57–1.66% and RPD values of 2.0–2.5, indicating a good approximate prediction capacity (% protein range 25.5–43.5%). Variables selected by iPLS obtained better regression models than variables selected by GA, based on the lower absolute error. Chemical maps displayed the heterogeneous protein distribution in single larvae and a batch of larvae. This manuscript demonstrates that NIR-HSI and chemometrics can be implemented as a fast screening method to estimate protein content in single BFS larvae.

Chia (Salvia hispanica) seeds degradation studied by fuzzy-c mean (FCM) and hyperspectral imaging and chemometrics - fatty acids quantification

Chia seeds are nutritious food because they have a high content of protein, polyunsaturated fatty acids (omega-3 and omega-6) and phenolic compounds. During storage, fatty acids are degraded, by oxidative and hydrolytic reactions, forming free fatty acids (FFA). In this work, we used Near Infrared Hyperspectral Imaging (NIR- HSI) and chemometrics to predict FFA acid value and fatty acids concentrations in chia seeds during storage. First, we explore the hyperspectral images by Fuzzy c-means (FCM), where it is possible to observe as chemical compounds are formed or degraded during storage. Second, PLSR models were developed to predict FFA value and fatty acids concentration. RPD values reached values higher then 2.0, indicating a good ability to estimate these chemical compounds, especially polyunsaturated fatty acids omega-3 and omega-6. Finally, NIR-hyperspectral imaging coupled with chemometrics allowed us to show the chemical degradation process of chia seeds during storage, mainly associated with polyunsaturated fatty acids degradation. Besides NIR-HSI showed to be a powerful technique to quantify the main fatty acids with high accuracy.

New approach to estimate macro and micronutrients in potato plants based on foliar spectral reflectance

Tissue testing used to assess the chemical contents in potato plants is considered laborious, time-consuming, destructive, and expensive. Ground-based sensors have been assessed to provide efficient information on nitrogen using leaf canopy reflectance. In potatoes, however, the main organ required for tissue testing is the petiole to estimate the elements of all nutrients. This research aims to assess whether there is a correlation between the chemical contents of potato petioles and leaf spectrum, and to examine whether the spectrum of dried or fresh leaves have higher correlation values. Petiole chemical contents of all elements were tested as a reference point. Leaves were split equally into dried and fresh groups for spectral analysis (400–2500 nm). Lasso Regression models were built to estimate concentrations in comparison to actual values. The performances of the model were tested using the Ratio of (standard error of) Prediction to (standard) Deviation (RPD). All elements showed reasonable to excellent RPD values except for sodium. All elements showed higher correlation in the dried testing mode except for nitrogen and potassium. The models showed that the most significant wavebands were in the visible and very near infrared range (400–1100 nm) for all macronutrients except magnesium and sulfur, while all micronutrients had the most significant wavebands in full range (400–2500 nm) with a common significant waveband at 1932 nm. The results show high potentials of a new approach to estimate potato plant elements based on foliar spectral reflectance.

Performance evaluation of variable selection methods coupled with partial least squares regression to determine the target component in solid samples

Variable selection can improve the robustness and prediction accuracy of partial least squares (PLS) regression models and decrease the calculation time by selecting the optimal subset of variables in multivariate calibration. In this study, the performance of two variable selection methods for wavelength interval and individual wavelength coupled with partial least squares regression are investigated by employing the experimental data of asiaticoside (AS) and madecassoside (MS) contents in centella total glucosides (CTG) and a public dataset of corn. The studied variable selection methods include interval partial least squares regression ( iPLS), backward interval partial least squares ( biPLS), synergy interval partial least squares regression ( siPLS), competitive adaptive reweighted sampling (CARS), uninformative variable elimination (UVE) and variable importance in projection (VIP). The results show that the implementation of variable selection methods improved the performance of the model compared with full-spectrum modeling. All variable selection methods improved the prediction of AS or MS contents in CTG. When latent variables for PLS models are less than 10 in the practical application, the RPD value of AS models by iPLS method is 7.5, and the RPD value of MS models by biPLS method is 2.9. The results of wavelength interval selection are better than individual wavelength selection, especially for iPLS and biPLS. The same results were obtained with the public data for moisture in corn, and the RPD value of biPLS model of moisture is 1.6. Therefore, the wavelength interval selection methods, such as iPLS or biPLS, are appropriate for improving the PLS model’s accuracy and robustness to determine the target components’ contents in solid samples.

Novel Application of NIR Spectroscopy for Non-Destructive Determination of 'Maraština' Wine Parameters.

This study investigates the colour and standard chemical composition of must and wines produced from the grapes from Vitis vinifera L., ‘Maraština’, harvested from 10 vineyards located in two different viticultural subregions of the Adriatic region of Croatia: Northern Dalmatia and Central and Southern Dalmatia. The aim was to explore the use of NIR spectroscopy combined with chemometrics to determine the characteristics of Maraština wines and to develop calibration models relating NIR spectra and physicochemical/colour data. Differences in the colour parameters (L*, a*, hue) of wines related to the subregions were confirmed. Colour difference (ΔE) of must vs. wine significantly differed for the samples from the Maraština grapes grown in both subregions. Principal component regression was used to construct the calibration models based on NIR spectra and standard physicochemical and colour data showing high prediction ability of the 13 studied parameters of must and/or wine (average R2 of 0.98 and RPD value of 6.8). Principal component analysis revealed qualitative differences of must and wines produced from the same grape variety but grown in different subregions.

Prediction of Heavy Metal Concentrations in Contaminated Sites from Portable X-ray Fluorescence Spectrometer Data Using Machine Learning

Portable X-ray fluorescence (pXRF) spectrometers provide simple, rapid, nondestructive, and cost-effective analysis of the metal contents in soils. The current method for improving pXRF measurement accuracy is soil sample preparation, which inevitably consumes significant amounts of time. To eliminate the influence of sample preparation on PXRF measurements, this study evaluates the performance of pXRF measurements in the prediction of eight heavy metals’ contents through machine learning algorithm linear regression (LR) and multivariate adaptive regression spline (MARS) models. Soil samples were collected from five industrial sites and separated into high-value and low-value datasets with pXRF measurements above or below the background values. The results showed that for Cu and Cr, the MARS models were better than the LR models at prediction (the MARS-R2 values were 0.88 and 0.78; the MARS-RPD values were 2.89 and 2.11). For the pXRF low-value dataset, the multivariate MARS models improved the pXRF measurement accuracy, with the R2 values improved from 0.032 to 0.39 and the RPD values increased by 0.02 to 0.37. For the pXRF high-value dataset, the univariate MARS models predicted the content of Cu and Cr with less calculation. Our study reveals that machine learning methods can better predict the Cu and Cr of large samples from multiple contaminated sites.

Near infrared reflectance spectroscopy as a tool to predict non-starch polysaccharide composition and starch digestibility profiles in common monogastric cereal feed ingredients

Carbohydrates are the main components of cereal grain ingredients and are comprised predominantly of starch which varies in the rate and extent of digestibility, and non-starch polysaccharides (NSP) from plant cell walls. These components are traditionally measured by chemical methods but there is interest in developing more rapid, cost effective predictions by use of near infrared reflectance spectroscopy (NIRS) technology. In this study, around 400 samples of wheat, corn, barley, wheat bran, oats, rye, corn germ, and sorghum were collected over several years to develop global NIR calibrations. Chemical analysis was performed for NSP, including details on solubility and individual constituent sugars, as well as separate determinations of cellulose and lignin. Total starch content and starch digestibility were characterized by the method of Englyst as rapidly (RDS) and slowly (SDS) digestible starch, and resistant starch (RS) determined as total starch content minus RDS and SDS. The same samples were scanned (400 to 2500 nm, every 0.5 nm) by two NIRS machines: a Foss model DS2500 and a Bruker Tango FT-NIR (Fourier transform near-infrared spectroscopy). The NIRS calibrations were developed by using the combined set of different cereal grains. The modified Partial Least Squares Regression method was employed to establish correlations between the reference chemical analysis and the collected NIRS spectra. The correlation between the NIRS and reference data for total and insoluble NSP and the main component sugars (arabinose, xylose, and glucose) were robust (R2 = >0.97) with a 1-VR > 0.96 and RPD values > 5.3. Similarly, the NIRS predicted lignin (R2 = 0.94; 1-VR = 0.93) and to a lesser extent cellulose (R2 = 0.70; 1-VR = 0.98) but still highly correlated with the reference data. The predicted total starch content was excellent (R2 = 0.99; 1-VR = 0.99). The NIRS prediction for RDS and SDS were lower in linearity (R2 = 0.81 and R2 = 0.97, respectively) but more than sufficient to provide an estimate of the starch digestion profile for many of the raw materials evaluated. In addition, the robust NIRS predictions for total starch content, RDS and SDS could then provide information to calculate the RS fraction of the ingredients. In conclusion, compared with the reference chemical data for this large set of cereal grains the global NIR calibrations developed provide a good prediction of NSP, monosaccharide sugars, lignin, cellulose and total starch content in cereal grains, and for the first time an NIR relationship with starch digestibility fractions in the grains has been shown.

Reflectance spectroscopy for assessing heavy metal pollution indices in mangrove sediments using XGBoost method and physicochemical properties

Single and composite pollution indices (SPI and CPI) of heavy metals are calculated with background heavy metal concentrations. They are efficient evaluation indicators of sediment pollution levels and could be directly compared among different regions and habitat types, however, they are rarely investigated in mangrove sediment with visible and near-infrared spectroscopy (VNIRS). With mangrove sediment samples collected from five regions and four habitat types, this study aimed to develop estimation models of heavy metal (Cr, Ni, Cu, Zn and Pb) pollution indices using XGBoost (extreme gradient boosting) method, and further to explore the effect of physicochemical properties (Fe, organic carbon (OC), clay and salinity) on the estimation accuracy. Kruskal-Walls test showed that regions and habitat types were significant factors affecting the six pollution indices. Compared with first derivative, second derivative and continuum removal reflectance, standard normal variate (SNV) reflectance was more accurate for estimating SPIs and CPI using XGBoost model. The XGBoost models with sensitive wavelengths of SNV reflectance outperformed that with whole spectrum (410 bands) in estimating the six pollution indices, and the ranking of estimation accuracy was CPI > SPICu > SPIZn > SPIPb > SPINi > SPICr. Compared with the XGBoost model with sensitive wavelengths alone, the model with additional physicochemical properties could increase R2 and RPD (the ratio of standard deviation to RMSE of validation in the test subset) values by 0.795–7.282% and 1.199–8.780%, and decrease RMSE values by 4.101–26.866% in estimating SPICr, SPINi and SPICu. Moreover, Fe and OC had greater effect on the estimation accuracy than clay and salinity. We conclude the integration of XGBoost and physicochemical properties holds potentials for accurately assessing heavy metal pollution indices in mangrove sediment with VNIRS, which can provide methodological basis for evaluating environment quality and ecological risk in mangrove ecosystem at the plot or landscape scale.

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.

Robustness and accuracy evaluation of moisture prediction model for black tea withering process using hyperspectral imaging

The rapid and non-destructive detection of moisture in withering leaves is an unsolved problem because the leaves are stacked together and have random orientation. To address this issue, this study aimed to establish more robust and accurate models. The performance of front side, back side and multi-region models were compared, and the front side model showed the worst transferability. Therefore, five effective wavelength (EW) selection algorithms were combined with a successive projection algorithm (SPA) to select EWs. It was found that the shuffled frog leaping algorithm (SFLA) combined with SPA was the best method for the front side model for moisture analyses. Based on the selected EWs, the extreme learning machine (ELM) became the model with the best self-verification result. Subsequently, moisture distribution maps of withering leaves were successfully generated. Considering the processing demand of withering leaves, local region models developed based on partial least squares and the SFLA-SPA method were applied to predict the moisture of withering leaves in the local and stacked region. The results showed that the RPD, Rcv and Rp values were above 1.6, 0.870 and 0.897, respectively. These results provide a useful reference for the non-destructive detection of moisture in withering leaves.