Improvement of Prediction Models for Nondestructive Detection of TVB-N Using Dual-Band Vis/NIR Spectroscopic Technique

2017 ◽  
Vol 60 (4) ◽  
pp. 1075-1082 ◽  
Author(s):  
Wenxiu Wang ◽  
Yankun Peng
Keyword(s):  
Light Source ◽  
Spectral Region ◽  
Near Infrared ◽  
Prediction Models ◽  
Great Influence ◽  
Nir Spectroscopy ◽  
Dual Band ◽  
Coefficient Of Determination ◽  
Band Selection ◽  
Nondestructive Detection

Abstract. This article discusses the influence of light source and band selection on prediction model performance. Two spectra acquisition systems for visible (Vis) and near-infrared (NIR) spectroscopy with a ring light source and a point light source were set up and compared based on the coefficient of variation (CV), signal-to-noise ratio (SNR), spectrum area change rate (ACR), and model results. Reflectance spectra of 61 pork samples were collected, and anomalous samples were eliminated by Monte Carlo method based on model cluster analysis. Partial least squares (PLS) models for total volatile basic nitrogen (TVB-N) based on a single spectral region (350-1100 nm or 1000-2500 nm) and a dual spectral region (350-2500 nm) were built to compare the influence of band choice. Based on the optimal chosen band, characteristic wavelengths were selected by competitive adaptive reweighted sampling (CARS), and a new PLS model was established. The results showed that spectra acquired with the ring light source had better stability and achieved optimal prediction models. The dual spectral region, which contained more comprehensive information on TVB-N, yielded better results than any single spectral region. Based on the dual-band spectra, a simplified PLS model using feature variables achieved a coefficient of determination in the prediction set (Rp2) of 0.8767 and standard error of prediction (SEP) of 2.8354 mg per 100 g. The results demonstrated that the choice of light source and modeling band had great influence on prediction results, and improvement of models would promote the application of Vis/NIR spectroscopy in on-line or portable detection. Keywords: Band selection, Light source, Nondestructive detection, Pork, TVB-N, Vis/NIR spectroscopy.

scholarly journals Rapid Nondestructive Simultaneous Detection for Physicochemical Properties of Different Types of Sheep Meat Cut Using Portable Vis/NIR Reflectance Spectroscopy System

Foods ◽  
2021 ◽  
Vol 10 (9) ◽  
pp. 1975
Author(s):  
Yanlei Li ◽  
Xiaochun Zheng ◽  
Dequan Zhang ◽  
Xin Li ◽  
Fei Fang ◽  
...  
Keyword(s):  
Physicochemical Properties ◽  
Shear Force ◽  
Near Infrared ◽  
Prediction Models ◽  
Simultaneous Detection ◽  
Nir Spectroscopy ◽  
Cooking Loss ◽  
Coefficient Of Determination ◽  
Sheep Meat ◽  
Different Types

The visible and near-infrared spectroscopy (Vis/NIRS) models for sheep meat quality evaluation using only one type of meat cut are not suitable for other types. In this study, a novel portable Vis/NIRS system was used to simultaneously detect physicochemical properties (pH, color L*, a*, b*, cooking loss, and shear force) for different types of sheep meat cut, including silverside, back strap, oyster, fillet, thick flank, and tenderloin cuts. The results show that the predictive abilities for all parameters could be effectively improved by spectral preprocessing. The coefficient of determination (Rp2) and residual predictive deviation (RPD) of the optimal prediction models for pH, L*, a*, b*, cooking loss, and shear force were 0.79 and 3.50, 0.78 and 2.28, 0.68 and 2.46, 0.75 and 2.62, 0.77 and 2.19, and 0.83 and 2.81, respectively. The findings demonstrate that Vis/NIR spectroscopy is a useful tool for predicting the physicochemical properties of different types of meat cut.

scholarly journals NIR SPECTROSCOPIC MODELS FOR PHENOTYPING WOOD TRAITS IN BREEDING PROGRAMS OF Eucalyptus benthamii

CERNE ◽  
2017 ◽  
Vol 23 (3) ◽  
pp. 367-375 ◽  
Author(s):  
Regiane Abjaud Estopa ◽  
Flaviana Reis Milagres ◽  
Ricardo Augusto Oliveira ◽  
Paulo Ricardo Gherardi Hein
Keyword(s):  
Near Infrared ◽  
Prediction Models ◽  
Genetic Selection ◽  
Basic Density ◽  
Chemical Properties ◽  
Nir Spectroscopy ◽  
Coefficient Of Determination ◽  
Breeding Programs ◽  
Total Carbohydrates ◽  
Wood Traits

ABSTRACT Wood characterization must be done in huge populations of Eucalyptus breeding programs in order to efficiently select potential trees. In this study, Eucalyptus benthamii wood was non-destructively characterized and the performance of near infrared (NIR) spectroscopy in estimating the wood basic density, lignin, extractive, glucose, xylan contents and total carbohydrates was evaluated. NIR models for wood traits were performed from 481 trees from E. benthamii progeny test (4-year-old) managed for pulp cultivated in Santa Catarina state, Southern Brazil. Increment cores were sampled for chemical and physical characterization in laboratory, as well as for NIR spectroscopy analyses. Three 350 samples were selected from PCA for model calibrations whereas 131 were reserved for independent test validation. The E. benthamii wood presented the standards required for Kraft pulp processing. The predictive NIR models showed satisfactory ability for estimating the chemical properties of wood. The prediction models for total lignin, extractive and xylan contents and total carbohydrates showed coefficients of determination of 0.53, 0.65; 0.36 and 0.53, with RPD values for these traits ranging from 1.3 to 2.3. The predictive model for basic density of wood and glucose presented low coefficient of determination (0.13 and 0.10). However, isn’t possible to use these models for ranking in genetic selection because there was no correlation. Therefore, NIR spectroscopy can potentially be applied in breeding programs, as it enables an early, non-destructive selection of trees with adequate physical and chemical properties for pulp production process.

scholarly journals Determination of Moisture Content and Basic Density of Poplar Wood Chips under Various Moisture Conditions by Near-Infrared Spectroscopy

2019 ◽  
Vol 65 (5) ◽  
pp. 548-555 ◽  
Author(s):  
Long Liang ◽  
Guigan Fang ◽  
Yongjun Deng ◽  
Zhixin Xiong ◽  
Ting Wu
Keyword(s):  
Moisture Content ◽  
Near Infrared ◽  
Prediction Models ◽  
Basic Density ◽  
Free Water ◽  
Nir Spectroscopy ◽  
Wood Chips ◽  
Coefficient Of Determination ◽  
Poplar Wood ◽  
Absorption Bands

AbstractThe potential of near-infrared (NIR) spectroscopy coupled with partial least-squares (PLS) regression was used to determine the moisture content and basic density of poplar wood chips. NIR spectra collected from the surface of wood chips were used to develop calibration models for moisture content and basic density predication, and various spectral preprocessing techniques were applied to improve the accuracy and robustness of the prediction models. The models were tested using totally independent sample sets and exhibited acceptable predictive performance for moisture content (coefficient of determination for prediction [R2p] = 0.98 and standard error of prediction [SEP] = 2.51 percent) and basic density (R2p = 0.87 and SEP = 17.61 kg m–3). In addition, the effect of moisture variations on prediction of basic density was investigated based on NIR spectra from wood chips under various moisture levels. The results demonstrated that broad absorption bands from water molecules, especially when free water exists in the cell lumen, overlap with informative signals related to wood properties and weaken the calibration relation between spectral features and basic density. Thus, maintaining wood chips in a low and even moisture state would help achieve reliable estimates of wood density by NIR analysis models.

scholarly journals Histamine Control in Raw and Processed Tuna: A Rapid Tool Based on NIR Spectroscopy

Foods ◽  
2021 ◽  
Vol 10 (4) ◽  
pp. 885
Author(s):  
Sergio Ghidini ◽  
Luca Maria Chiesa ◽  
Sara Panseri ◽  
Maria Olga Varrà ◽  
Adriana Ianieri ◽  
...  
Keyword(s):  
Root Mean Square ◽  
Near Infrared ◽  
Nir Spectroscopy ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression ◽  
Coefficient Of Determination ◽  
Mean Square ◽  
Tuna Fish ◽  
Mean Square Errors

The present study was designed to investigate whether near infrared (NIR) spectroscopy with minimal sample processing could be a suitable technique to rapidly measure histamine levels in raw and processed tuna fish. Calibration models based on orthogonal partial least square regression (OPLSR) were built to predict histamine in the range 10–1000 mg kg−1 using the 1000–2500 nm NIR spectra of artificially-contaminated fish. The two models were then validated using a new set of naturally contaminated samples in which histamine content was determined by conventional high-performance liquid chromatography (HPLC) analysis. As for calibration results, coefficient of determination (r2) > 0.98, root mean square of estimation (RMSEE) ≤ 5 mg kg−1 and root mean square of cross-validation (RMSECV) ≤ 6 mg kg−1 were achieved. Both models were optimal also in the validation stage, showing r2 values > 0.97, root mean square errors of prediction (RMSEP) ≤ 10 mg kg−1 and relative range error (RER) ≥ 25, with better results showed by the model for processed fish. The promising results achieved suggest NIR spectroscopy as an implemental analytical solution in fish industries and markets to effectively determine histamine amounts.

Prediction of Fatty Acid Content in Beef Neck Lean by near Infrared Reflectance Analysis

1998 ◽  
Vol 6 (1) ◽  
pp. 229-234 ◽  
Author(s):  
William R. Windham ◽  
W.H. Morrison
Keyword(s):  
Fatty Acids ◽  
Fatty Acid ◽  
Near Infrared ◽  
Fatty Acid Content ◽  
Reference Method ◽  
Nir Spectroscopy ◽  
Coefficient Of Determination ◽  
Longissimus Dorsi ◽  
Near Infrared Reflectance ◽  
Total Fatty Acids

Near infrared (NIR) spectroscopy in the prediction of individual and total fatty acids of bovine M. Longissimus dorsi neck muscles has been studied. Beef neck lean was collected from meat processing establishments using advanced meat recovery systems and hand-deboning. Samples ( n = 302) were analysed to determine fatty acid (FA) composition and scanned from 400 to 2498 nm. Total saturated and unsaturated FA values ranged from 43.2 to 62.0% and 38.3 to 56.2%, respectively. Results of partial least squares (PLS) modeling shown reasonably accurate models were attained for total saturate content [standard error of performance ( SEP = 1.10%); coefficient of determination on the validation set ( r2 = 0.77)], palmitic ( SEP = 0.94%; r2 = 0.69), unsaturate ( SEP = 1.13%; r2 = 0.77), and oleic ( SEP = 0.97; r2 = 0.78). Prediction of other individual saturated and unsaturated FAs was less accurate with an r2 range of 0.10 to 0.53. However, the sum of individual predicted saturated and unsaturated FA was acceptable compared with the reference method ( SEP = 1.10 and 1.12%, respectively). This study shows that NIR can be used to predict accurately total fatty acids in M. Longissimus dorsi muscle.

scholarly journals Machine Learning Applied to Near-Infrared Spectra for Chicken Meat Classification

2018 ◽  
Vol 2018 ◽  
pp. 1-12 ◽  
Author(s):  
Sylvio Barbon ◽  
Ana Paula Ayub da Costa Barbon ◽  
Rafael Gomes Mantovani ◽  
Douglas Fernandes Barbin
Keyword(s):  
Near Infrared ◽  
Prediction Models ◽  
Nir Spectroscopy ◽  
Quality Parameters ◽  
Chicken Meat ◽  
Poultry Meat ◽  
Spectral Information ◽  
Support Vector ◽  
Near Infrared Spectra ◽  
Meat Samples

Identification of chicken quality parameters is often inconsistent, time-consuming, and laborious. Near-infrared (NIR) spectroscopy has been used as a powerful tool for food quality assessment. However, the near-infrared (NIR) spectra comprise a large number of redundant information. Determining wavelengths relevance and selecting subsets for classification and prediction models are mandatory for the development of multispectral systems. A combination of both attribute and wavelength selection for NIR spectral information of chicken meat samples was investigated. Decision Trees and Decision Table predictors exploit these optimal wavelengths for classification tasks according to different quality grades of poultry meat. The proposed methodology was conducted with a support vector machine algorithm (SVM) to compare the precision of the proposed model. Experiments were performed on NIR spectral information (1050 wavelengths), colour (CIEL∗a∗b∗, chroma, and hue), water holding capacity (WHC), and pH of each sample analyzed. Results show that the best method was the REPTree based on 12 wavelengths, allowing for classification of poultry samples according to quality grades with 77.2% precision. The selected wavelengths could lead to potential simple multispectral acquisition devices.

Application of near Infrared Spectroscopy On-Combine in Corn Grain Breeding

2005 ◽  
Vol 13 (2) ◽  
pp. 69-75 ◽  
Author(s):  
Roland Welle ◽  
Willi Greten ◽  
Thomas Müller ◽  
Gary Weber ◽  
Hartwig Wehrmann
Keyword(s):  
Grain Yield ◽  
Dry Matter ◽  
Near Infrared ◽  
Reference Method ◽  
Nir Spectroscopy ◽  
Coefficient Of Determination ◽  
Successful Implementation ◽  
Grain Moisture ◽  
Yield Determination ◽  
Corn Grain

Improving maize ( Zea mays L.) grain yield and agronomic properties are major goals for corn breeders in northern Europe. In order to facilitate field grain yield determination we measured corn grain moisture content with near infrared (NIR) spectroscopy directly on a harvesting machine. NIR spectroscopy, in combination with harvesting, significantly improved quality and speed of yield determination within the very narrow harvest time window. Moisture calibrations were developed with 2117 samples from the 2001 to 2003 crop seasons using six diode array spectrometers mounted on combines. These models were derived from databases containing spectra from all instruments. Spectrometer-specific calibrations cannot be used to predict samples measured on other instruments of the same type. Standard error of cross-validation ( SECV) and coefficient of determination ( R2) were 0.56 and 0.99%, respectively. Moisture standard errors of prediction ( SEPs) for the six instruments, using varying independent sample sets from the 2004 harvest, ranged between 0.59% and 0.99% with R2 values between 0.92 to 0.98. The six instruments produced the same dry matter predictions on a common sample set as indicated by high R2 and low biases among them, hence there was no need to apply specific standardisation algorithms. Moisture NIR spectroscopy determinations were significantly more precise than those obtained using the reference method. Analysis of variance revealed low least significant differences and high heritabilities. High precision and heritability demonstrate successful implementation of on-combine NIR spectroscopy for routine dry matter (yield) measurements.

scholarly journals Predicting terpene content in dried conifer shoots using near infrared spectroscopy

2020 ◽  
Vol 28 (5-6) ◽  
pp. 308-314
Author(s):  
Emilie Champagne ◽  
Michaël Bonin ◽  
Alejandro A Royo ◽  
Jean-Pierre Tremblay ◽  
Patricia Raymond
Keyword(s):  
Gas Chromatography ◽  
Picea Glauca ◽  
Near Infrared ◽  
Nir Spectroscopy ◽  
Calibration Method ◽  
Wide Distribution ◽  
Coefficient Of Determination ◽  
Calibration Model ◽  
Freeze Dried ◽  
Terpene Content

Terpenes are phytochemicals found in multiple plant genera, especially aromatic herbs and conifers. Terpene content quantification is costly and complex, requiring the extraction of oil content and gas chromatography analyses. Near infrared (NIR) spectroscopy could provide an alternative quantitative method, especially if calibration can be developed with the spectra of dried plant material, which are easier and faster to acquire than oil-based spectra. Here, multispecies NIR spectroscopy calibrations were developed for total terpene content (mono- and sesquiterpenes) and for specific terpenes (α-pinene, β-pinene and myrcene) with five conifers species ( Picea glauca, Picea rubens, Pinus resinosa, Pinus strobus and Thuja occidentalis). The terpene content of fresh shoot samples was quantified with gas chromatography. The NIR spectra were measured on freeze-dried samples (n = 137). Using a subset of the samples, modified partial least squares regressions of total terpene and the three individual terpenes content were generated as a functions of the NIR spectra. The standard errors of the internal cross-validations (values between 0.25 and 2.28) and the ratio of prediction to deviation ratios (RPD values between 2.20 and 2.38) indicate that all calibrations have similar accuracy. The independent validations, however, suggest that the calibrations for total terpene and α-pinene content are more accurate (respective coefficient of determination: r2 = 0.85 and 0.82). In contrast, calibrations for β-pinene and myrcene had a low accuracy (respectively: r2 = 0.62 and 0.08), potentially because of the low concentration of these terpenes in the species studied. The calibration model fits (i.e., r2) are comparable to previously published calibration using the spectra of dried shoot samples and demonstrate the potential of this method for terpenes in conifer samples. The calibration method used could be useful in several other domains (e.g. seedling breeding program, industrial), because of the wide distribution of terpenes and especially of pinenes.

Rapid analysis of chemical composition in intact and milled rice cookies using near infrared spectroscopy

2017 ◽  
Vol 25 (5) ◽  
pp. 330-337 ◽  
Author(s):  
Latthika Wimonsiri ◽  
Pitiporn Ritthiruangdej ◽  
Sumaporn Kasemsumran ◽  
Nantawan Therdthai ◽  
Wasaporn Chanput ◽  
...  
Keyword(s):  
Chemical Composition ◽  
Near Infrared ◽  
Wheat Gluten ◽  
Nir Spectroscopy ◽  
Rice Flour ◽  
Coefficient Of Determination ◽  
Gluten Free ◽  
Least Squares Regression ◽  
Milled Rice ◽  
Control Programs

This study has investigated the potential of near infrared (NIR) spectroscopy to predict the content of moisture, protein, fat and gluten in rice cookies in different sample forms (intact and milled samples). Gluten-free (n = 48) and gluten (n = 48) rice cookies were formulated with brown and white rice flours in which butter was substituted with fat replacer at 0, 15, 30 and 45%. With regard to gluten cookies, rice flour was substituted with wheat gluten at 1, 3 and 5%. Partial least squares regression modeling produced models with coefficient of determination (R2) values greater than 0.88 from NIR spectra of intact samples and greater than 0.92 for milled samples. These models were able to predict the four components with a ratio of prediction to deviation greater than 2.7 and 3.8 in intact and milled samples, respectively. The results suggest that the models obtained from the intact samples can be successfully applied for chemical composition of rice cookies and are reliable enough use for potential quality control programs.

Prediction of Yellow-Poplar (Liriodendron Tulipifera) Veneer Stiffness and Bulk Density Using near Infrared Spectroscopy and Multivariate Calibration

2008 ◽  
Vol 16 (5) ◽  
pp. 487-496 ◽  
Author(s):  
Oluwatosin Emmanuel Adedipe ◽  
Ben Dawson-Andoh
Keyword(s):  
Bulk Density ◽  
Spectral Range ◽  
Predictive Models ◽  
Spectral Region ◽  
Near Infrared ◽  
Multivariate Calibration ◽  
Nir Spectroscopy ◽  
Yellow Poplar ◽  
First Derivative ◽  
Calibration Models

This study investigated the feasibility of using near infrared (NIR) spectroscopy and multivariate calibration to predict bulk density and stiffness of 3.2 mm thick yellow poplar veneer strips. Full-range (800–2500 nm) raw NIR spectra and spectra pre-treated using the first derivative method, along with spectra from three other different wavelength windows of 1200–2400 nm, 1800–2400 nm and 1400–2000 nm were regressed against the bulk density (kg m−3) values and the dynamic modulus of elasticity (stiffness; GPa) of the veneers using the projection to latent structures (PLS) method to develop calibration models. All predictive models developed performed well in the prediction of bulk density and stiffness of new test samples that were not included in the calibration models. R2 values ranged from 0.67-0.78 and 0.56-0.72, respectively, for bulk density and stiffness. There was significant improvement in models developed with first derivative spectra over models developed with raw spectra. The models developed using the first derivative used fewer latent variables to achieve predictive models with higher R2 values, lower root mean square errors of prediction (RMSEP) and standard errors of prediction (SEP). Models developed using the full NIR spectral range (800–2500 nm) and the NIR spectral region of 1200–2400 nm performed better than models developed using the restricted NIR wavelength regions of 1800–2400 nm and 1400–2000 nm. However, there was no clear distinction between models developed using the full NIR spectral range and the NIR spectral region of 1200–2400 nm. Overall, models developed with the first derivative pre-processed spectra using the whole NIR spectrum performed best in predictability. The results of this study show the potential of using multivariate data analysis coupled with NIR spectroscopy for on-line sorting and assessment of veneer stiffness prior to the lay-up process in the manufacturing of veneer-based engineered wood products such as plywood, Parallam and laminated veneer lumber.

Sign in / Sign up

Export Citation Format

Share Document