scholarly journals Deep Spectral-Spatial Features of Near Infrared Hyperspectral Images for Pixel-Wise Classification of Food Products

Sensors ◽  
2020 ◽  
Vol 20 (18) ◽  
pp. 5322
Author(s):  
Hongyan Zhu ◽  
Aoife Gowen ◽  
Hailin Feng ◽  
Keping Yu ◽  
Jun-Li Xu
Keyword(s):  
Near Infrared ◽  
Principal Component ◽  
Food Products ◽  
Partial Least Square ◽  
Least Square ◽  
Support Vector ◽  
White Stripe ◽  
Practical Applications ◽  
Spatial Features

Hyperspectral imaging (HSI) emerges as a non-destructive and rapid analytical tool for assessing food quality, safety, and authenticity. This work aims to investigate the potential of combining the spectral and spatial features of HSI data with the aid of deep learning approach for the pixel-wise classification of food products. We applied two strategies for extracting spatial-spectral features: (1) directly applying three-dimensional convolution neural network (3-D CNN) model; (2) first performing principal component analysis (PCA) and then developing 2-D CNN model from the first few PCs. These two methods were compared in terms of efficiency and accuracy, exemplified through two case studies, i.e., classification of four sweet products and differentiation between white stripe (“myocommata”) and red muscle (“myotome”) pixels on salmon fillets. Results showed that combining spectral-spatial features significantly enhanced the overall accuracy for sweet dataset, compared to partial least square discriminant analysis (PLSDA) and support vector machine (SVM). Results also demonstrated that spectral pre-processing techniques prior to CNN model development can enhance the classification performance. This work will open the door for more research in the area of practical applications in food industry.

Classification of Automobile Lubricant by Near-Infrared Spectroscopy Combined with Machine Classification

2011 ◽  
Vol 460-461 ◽  
pp. 667-672
Author(s):  
Yun Zhao ◽  
Xing Xu ◽  
Yong He
Keyword(s):  
Near Infrared ◽  
Prediction Models ◽  
Nir Spectroscopy ◽  
Partial Least Square ◽  
Least Square ◽  
Support Vector ◽  
Training Set ◽  
Test Set ◽  
Svm Model

The main objective of this paper is to classify four kinds of automobile lubricant by near-infrared (NIR) spectral technology and to observe whether NIR spectroscopy could be used for predicting water content. Principle component analysis (PCA) was applied to reduce the information from the spectral data and first two PCs were used to cluster the samples. Partial least square (PLS), least square support vector machine (LS-SVM), and Gaussian processes classification (GPC) were employed to develop prediction models. There were 120 samples for training set and test set. Two LS-SVM models with first five PCs and first six PCs were built, respectively, and accuracy of the model with five PCs is adequate with less calculation. The results from the experiment indicate that the LS-SVM model outperforms the PLS model and GPC model outperforms the LS-SVM model.

scholarly journals Rapid Classification Of Agricultural Products Based On Their Electro-Optic Properties Using Near Infrared Reflectance And Chemometrics

2015 ◽  
Vol 8 (1) ◽  
pp. 10-18
Author(s):  
Agus A. Munawar
Keyword(s):  
Near Infrared ◽  
Scatter Correction ◽  
Principal Component ◽  
Partial Least Square ◽  
Least Square ◽  
Agricultural Products ◽  
Near Infrared Reflectance ◽  
Electro Optic ◽  
Non Destructive

Abstract. Near infrared technology have been widely applied in many fields, including agriculture especially in sorting and grading process. The advantage of this technology: simple sample preparation, rapid, effective and non-destructive. The main objective of this study is to evaluate the feasibility of NIR technology in classifying several agricultural products based on their electro-optic properties. NIR diffuse reflectance spectra of apples, bananas, mangoes, garlics, tomatoes, green grapes, red grapes and oranges were acquired in wavelength range of 1000-2500 nm with gradual increment of 2 nm. Chemometrics methods were applied in combination with NIR spectra data. Classification was performed by applying principal component analysis (PCA) followed by non-iterative partial least square (NIPALS) cross validation. The results showed that NIR and chemometrics was able to differentiate and classify these agricultural products with two latent variables (2 PCs) and total explained variance of 97% (88% PC1 and 9% PC2). Furthermore, it also showed that multiplicative scatter correction (MSC) was found to be effective spectra correction or enhancement method and increased classification accuracy and robustness. It may conclude that NIR technology combined with chemometrics was feasible to apply as a rapid and non-destructive method for sorting and grading agricultural products. Rapid Classification Of Agricultural Products Based On Their Electro-Optic Properties Using Near Infrared Reflectance And ChemometricsAbstract. Aplikasi teknologi near infra red (NIR) telah digunakan dalam banyak bidang, termasuk untuk bidang pertanian terutama pada proses sortasi dan grading. Keunggulan metode ini antara lain : rapid, efektif, simultan dan tanpa merusak objek yang dikaji. Tujuan utama dari studi ini adalah untuk mengkaji potensi NIR dalam mengklasifikasi beberapa produk pertanian berdasarkan karakteristik sifat elektro-optik dari produk tersebut. Spektrum NIR pada panjang gelombang 1000 – 2500 nm dengan increment 2 nm diakuisisi untuk produk pertanian : apel, pisang, manga, bawang putih, tomat, anggur hijau, anggur merah dan jeruk. Metode chemo metrics digunakan dalam studi ini untuk dikombinasikan dengan spektrum NIR. Klasifikasi produk pertanian dilakukan dengan menerapkan metode principal component analysis (PCA) yang disertai dengan metode non-iterative partial least square (NIPALS) cross validation. Hasil studi menunjukkan bahwa kombinasi NIR dan chemo metrics mampu membedakan dan mengklasifikasi produk pertanian tersebut dengan menggunakan dua latent variable pada PCA (2 PCs) dengan total explained variance 97% (88% PC1 dan 9% PC2). Selain itu, dari studi ini juga didapatkan bahwa perbaikan data spectrum dengan metode multiplicative scatter correction (MSC) sebelum klasifikasi mampu meningkatkan akurasi hasil klasifikasi. Secara umum, dapat disimpulkan bahwa teknologi NIR dan chemo metrics dapat dijadikan sebagai metode yang efektif untuk sortasi dan atau grading produk pertanian.

Rapid Determining Contents of the Rhubarb Anthraquinones Compounds by Support Vector Machines Modeling based on Near Infrared Spectra

2020 ◽  
Vol 16 ◽  
Author(s):  
Linqi Liu ◽  
JInhua Luo ◽  
Chenxi Zhao ◽  
Bingxue Zhang ◽  
Wei Fan ◽  
...  
Keyword(s):  
Infrared Spectra ◽  
Near Infrared ◽  
Mean Squared Error ◽  
Rapid Determination ◽  
Partial Least Square ◽  
Least Square ◽  
Support Vector ◽  
Near Infrared Spectra ◽  
Aloe Emodin ◽  
Relative Differences

BACKGROUND: Measuring medicinal compounds to evaluate their quality and efficacy has been recognized as a useful approach in treatment. Rhubarb anthraquinones compounds (mainly including aloe-emodin, rhein, emodin, chrysophanol and physcion) are its main effective components as purgating drug. In the current Chinese Pharmacopoeia, the total anthraquinones content is designated as its quantitative quality and control index while the content of each compound has not been specified. METHODS: On the basis of forty rhubarb samples, the correlation models between the near infrared spectra and UPLC analysis data were constructed using support vector machine (SVM) and partial least square (PLS) methods according to Kennard and Stone algorithm for dividing the calibration/prediction datasets. Good models mean they have high correlation coefficients (R2) and low root mean squared error of prediction (RMSEP) values. RESULTS: The models constructed by SVM have much better performance than those by PLS methods. The SVM models have high R2 of 0.8951, 0.9738, 0.9849, 0.9779, 0.9411 and 0.9862 that correspond to aloe-emodin, rhein, emodin, chrysophanol, physcion and total anthraquinones contents, respectively. The corresponding RMSEPs are 0.3592, 0.4182, 0.4508, 0.7121, 0.8365 and 1.7910, respectively. 75% of the predicted results have relative differences being lower than 10%. As for rhein and total anthraquinones, all of the predicted results have relative differences being lower than 10%. CONCLUSION: The nonlinear models constructed by SVM showed good performances with predicted values close to the experimental values. This can perform the rapid determination of the main medicinal ingredients in rhubarb medicinal materials.

scholarly journals A fast and low-cost approach to quality control of alcohol-based hand sanitizer using portable near infrared spectrometer and chemometrics

2021 ◽  
pp. 096703352098731
Author(s):  
Adenilton C da Silva ◽  
Lívia PD Ribeiro ◽  
Ruth MB Vidal ◽  
Wladiana O Matos ◽  
Gisele S Lopes
Keyword(s):  
Quality Control ◽  
Discriminant Analysis ◽  
Near Infrared ◽  
Nir Spectroscopy ◽  
Principal Component ◽  
Partial Least Square ◽  
Least Square ◽  
Chemical Components ◽  
Linear Discriminant ◽  
Hand Sanitizer

The use of alcohol-based hand sanitizers is recommended as one of several strategies to minimize contamination and spread of the COVID-19 disease. Current reports suggest that the virucidal potential of ethanol occurs at concentrations close to 70%. Traditional methods of verifying the ethanol concentration in such products invite potential errors due to the viscosity of chemical components or may be prohibitively expensive to undertake in large demand. Near infrared (NIR) spectroscopy and chemometrics have already been used for the determination of ethanol in other matrices and present an alternative fast and reliable approach to quality control of alcohol-based hand sanitizers. In this study, a portable NIR spectrometer combined with classification chemometric tools, i.e., partial least square discriminant analysis (PLS–DA) and linear discriminant analysis with successive algorithm projection (SPA–LDA) were used to construct models to identify conforming and non-conforming commercial and laboratory synthesized hand sanitizer samples. Principal component analysis (PCA) was applied in an exploratory data study. Three principal components accounted for 99% of data variance and demonstrate clustering of conforming and non-conforming samples. The PLS–DA and SPA–LDA classification models presented 77 and 100% of accuracy in cross/internal validation respectively and 100% of accuracy in the classification of test samples. A total of 43% commercial samples evaluated using the PLS–DA and SPA–LDA presented ethanol content non-conforming for hand sanitizer gel. These results indicate that use of NIR spectroscopy and chemometrics is a promising strategy, yielding a method that is fast, portable, and reliable for discrimination of alcohol-based hand sanitizers with respect to conforming and non-conforming ethanol concentrations.

scholarly journals Prediction of macro- and microelements content in Croatian common bean landraces (Phaseolus vulgaris L.) by NIR spectroscopy

Poljoprivreda ◽  
2019 ◽  
Vol 25 (1) ◽  
pp. 48-55
Author(s):  
Marina Vranić ◽  
Marko Petek ◽  
Krešimir Bošnjak ◽  
Boris Lazarević ◽  
Klaudija Carović Stanko
Keyword(s):  
Phaseolus Vulgaris ◽  
Common Bean ◽  
Near Infrared ◽  
Model Development ◽  
Nir Spectroscopy ◽  
Principal Component ◽  
Partial Least Square ◽  
Least Square ◽  
Potassium Oxide ◽  
Phaseolus Vulgaris L

In this study, near-infrared spectroscopy (NIRS) was used to predict the contents of essential macro- and microelements in common bean (Phaseolus vulgaris L.) accessions of most widespread Croatian landraces. Total of 175 samples were used for the model development by modified partial least square (MPLS), principal component regression (PCR) and partial least square (PLS) techniques. Based on the coefficients of determination (R2), standard error of calibration (SEC) and error of prediction (SEP) the models developed were (i) nearly applicable for nitrogen (N) (0.89, 0.12 and 0.45 respectively), (ii) poor for iron (Fe), cinc (Zn), potassium oxide (K2O) and potassium (K), (iii) usable for phosphorus pentoxide (P2O5), phosphorus (P), phytic acid (PA) and manganese (Mn). The MPLS regression statistics suggested the most accurate models developed comparing with PLS and PCR. It was concluded that a wider set of common bean samples needs to be used for macro- and microelements prediction by NIRS.

scholarly journals Prediction of Soil Texture Distributions by using PLSR and Reflectance Spectroscopy

2019 ◽  
Vol 8 (3) ◽  
pp. 7876-7881
Keyword(s):  
Statistical Analysis ◽  
Soil Texture ◽  
Near Infrared ◽  
Diffuse Reflectance Spectroscopy ◽  
Principal Component ◽  
Reflectance Spectroscopy ◽  
Soil Samples ◽  
Partial Least Square ◽  
Least Square ◽  
Partial Least Square Regression

The texture of soil i.e. Sand, Silt and Clay are the most important physical properties of soil for agricultural management. In the agricultural practices to increase the productivity of soil, moisture-holding capacity, aeration and to support the agronomic decisions the knowledge of soil texture is an essential task. For this purpose, the present research gives better results and fast acquisition of soil information with the use of Visible and Near Infrared (Vis- NIR) Diffuse Reflectance Spectroscopy. A total of 30 soil samples from two different locations from Aurangabad, Maharashtra, India were collected and analyzed for soil texture. To detect the soil texture the Vis-NIR DRS has shown levels of accurate results compared to the traditional laboratory method with less time, cost and effort. To measure the reflectance of soil the ASD FieldSpec4 Spectroradiometer (350-2500nm) was used. By the observation of captured spectra by using Spectroradiometer it showed that on the basis of different textural classes the soil samples could be spectrally separable. For database collection and pre-processing, we have used RS3 and ViewSpec Pro software respectively. The statistical analysis by using the combination of Principal Component Analysis (PCA) and Partial Least Square Regression method gives accurate results. To determine the texture of soil sample thirteen features were calculated. The main goal of this research was to determine the soil texture by using statistical methods and to test the performance of VNIR-SWIR reflectance spectroscopy by using the ASD FieldSpec4 Spectroradiometer for estimation of the texture of the soil. The results showed that R2 = 0.99 gives maximum accuracy for clay content and R2 = 0.988 for silt content and R2 = 0.989 for sand. The Root Mean Square Values (RMSE) for clay, silt, and sand are 0.02392, 0.02399 and 0.02289 respectively. With the use of reflectance spectroscopy and statistical analysis by using regression models we can determine the soil properties accurately in very less time.

Visible-near infrared spectroscopy for detection of blood in sheep faeces

2020 ◽  
Vol 28 (5-6) ◽  
pp. 255-266 ◽  
Author(s):  
Elise A Kho ◽  
Jill N Fernandes ◽  
Andrew C Kotze ◽  
Glen P Fox ◽  
Maggy Lord ◽  
...  
Keyword(s):  
Support Vector Machine ◽  
Infrared Spectroscopy ◽  
Near Infrared Spectroscopy ◽  
Near Infrared ◽  
Mean Squared Error ◽  
Partial Least Square ◽  
Least Square ◽  
Support Vector ◽  
Root Mean Squared Error ◽  
Squared Error

Heavy infestations of the blood-sucking gastrointestinal nematodes, Haemonchus contortus can cause severe anaemia in sheep and leakage of blood into the faeces, leading to morbidity and mortality. Early and accurate diagnosis of infections is critical for timely treatment of sheep, minimizing production and sheep welfare impacts. In pursuit of a quick and easy measure of H. contortus infections, we investigated the use of portable visible near infrared spectrometers for detecting the presence of haemoglobin in sheep faeces as an indicator of H. contortus infection. Calibration models built within the 400–600 nm region by partial least square regression resulted in acceptable prediction accuracies (r 2 p > 0.70 and root mean squared error of prediction <2.64 µg Hb mg−1 faeces) for haemoglobin quantification using two spectrometers. The prediction results from support vector machine regression further improved the prediction of haemoglobin in moist sheep faeces (r 2 p > 0.87 and root mean squared error of prediction <2.00 µg haemoglobin mg−1 faeces). Based on a threshold for anthelmintic treatment of 3 µg Hb mg−1 faeces, both the partial least square and support vector machine models showed high sensitivity (89%) and high specificity (>77%). The specificity of the prediction model for detecting haemoglobin in sheep faeces may be improved by adding more variations in faecal composition into the calibration model. Our success in detecting haemoglobin in sheep faeces, following minimal sample preparation, suggests that with further development, vis–near infrared spectroscopy can provide a sensitive and convenient method for on-farm diagnosis of H. contortus infections.

scholarly journals Prediction of Soil-Available Potassium Content with Visible Near-Infrared Ray Spectroscopy of Different Pretreatment Transformations by the Boosting Algorithms

2020 ◽  
Vol 10 (4) ◽  
pp. 1520
Author(s):  
Xiu Jin ◽  
Shaowen Li ◽  
Wu Zhang ◽  
Juanjuan Zhu ◽  
Jia Sun
Keyword(s):  
Regression Models ◽  
Near Infrared ◽  
Potassium Concentration ◽  
Partial Least Square ◽  
Least Square ◽  
Support Vector ◽  
Standard Normal Variate ◽  
Standard Normal ◽  
Boosting Algorithms ◽  
Near Infrared Ray

The application of visible near-infrared (VIS-NIR) analysis technology to quantify the nutrients in soil has been widely recognized. It is important to improve the performance of regression models that can predict the soil-available potassium concentration. This study collected soil samples from southern Anhui, China, and concentrated on the modelling methods by using 29 pretreatment methods. The results show that a combination of three methods, Savitzky–Golay, standard normal variate, and dislodge tendency, exhibited better stability than others because it was the most capable of achieving levels A and B of the ratio of performance of deviation. The boosting algorithms that form an ensemble of multiple weak predictors exhibited better performance than partial least square (PLS) regression and support vector regression (SVR) for the prediction of soil-available potassium. These regression models could be employed to precisely predict the soil-available potassium concentration.

Sign in / Sign up

Export Citation Format

Share Document