摘要:
A variety of recovered-edible-oil identification models were established by using Raman combined with near-infrared spectroscopy (NIR). Eight types of 156 edible vegetable oil samples were collected to acquire their Raman and NIR spectra. The spectral data were processed for modeling. The preprocessing methods for the Raman spectra included the moving average method (11 points), adaptive iterative reweighted-penalty least squares method, and the normalization method based on the intensity of the characteristic peak at 1454 cm-1 (MA11-airPLS-Nor). The preprocessing method for the NIR spectra was the standard normal variable transformation algorithm combined with a detrending technique (SNV_DT). The Raman and NIR spectra were fused at the feature level by using independently the serial fusion and wavelet fusion approaches. The results showed that with the serial-fusion-and wavelet-fusion-based models, the identification of recovered oils can be achieved very rapidly. Furthermore, the comprehensive performances of the models based on fused Raman and NIR data were better than those of models based on separate Raman or NIR data.
作者机构:
[Xiao Zheng; Yaru Yu; Qingsong Luo; Qiang Xu; Yang Chen] School of Mechanical Engineering, Wuhan Polytechnic University, China
会议名称:
2018 2nd International Conference on Electronic Information Technology and Computer Engineering (EITCE 2018)(2018第二届电子信息技术与计算机工程国际会议)(EITCE2018)
会议时间:
2018-10-12
会议地点:
上海
会议论文集名称:
2018 2nd International Conference on Electronic Information Technology and Computer Engineering (EITCE 2018)(2018第二届电子信息技术与计算机工程国际会议)(EITCE2018)论文集
摘要:
To identify the type of edible oil, we proposed a near-infrared (NIR) spectral analysis method based on the contents of four characteristic components: iodine value, palmitic acid, oleic acid, and linoleic acid. Built support vector machine qualitative models that can identify eight kinds of oil. The four characteristic component values of 129 oil samples from 8 kinds of oil were collected and detected. Established three SVC identification models by using three parameter optimization methods including genetic algorithm, grid search and particle swarm optimization. The results showed that the prediction sets' accuracy rates of all the three models were up to 100%. Especially, both the accuracy rates of the correction and prediction sets of the particle-swarm-optimization-support-vector-machine classification (PSO-SVC) model reached 100%. The results indicate that it is effective and feasible to use the contents of characteristic components to identify the type of edible oil, and this method is fast, convenient, and accurate.
会议名称:
2nd International Conference on Applied Mathematics, Simulation and Modelling (AMSM)
会议时间:
AUG 06-07, 2017
会议地点:
Adv Sci & Ind Res Ctr, Phuket, THAILAND
会议主办单位:
Adv Sci & Ind Res Ctr
会议论文集名称:
DEStech Transactions on Engineering and Technology Research
关键词:
Near infrared;Saponification value;Parameter optimization;SVR
摘要:
Regression prediction methods of the saponification value of edible oils,which were based on different parameter optimization algorithms combined with near infrared spectroscopy(NIR) were studied.The
摘要:
The purpose of this study is to conduct qualitative analysis on the adulteration in peanut oil by combining data fusion of Raman and near infrared (NIR) spectral characteristics with chemometrics methods. With laser Raman and NIR spectrometer, the spectra of 134 adulterated oil samples and 24 pure peanut oil were collected. The spectra data of Raman and NIR were preprocessed. Competitive adaptive reweighted sampling(CARS) were used to extract the characteristic wavelengths of the spectra data. Combining data fusion technique and partial least squares linear discriminant analysis (PLS-LDA) method, the Ram-PLS-LDA model, NIR-PLS-LDA model and Ram-NIR-PLS-LDA model were established by using the obtained feature layer data. The calibration set and prediction set accuracy of the SG9-airPLS-Nor-CARS-SNV_DTCARS-PLS-LDA model are 100%. According to the analysis, the prediction accuracy of Ram-NIR-PLS-LDA model is better than that of single spectral model, data fusion technology can enhance the ability to identify the model, which is conducive to practical application. It shows that the two kinds of spectra are complementary, and the using of spectral analysis and data fusion technology has great application value in the identification of edible oil.
摘要:
An approach based on multi-source spectra data fusion for identification of edible oil is proposed. A qualitative model based on fusion of Raman spectra and near-infrared spectroscopy (Raman-NIR) was established and compared with conventional single-spectra model. The spectra data was pre-processed using the moving average method (MA11), the Savitzky-Golay method (SG9), the adaptive iteratively reweighted penalized least squares method (airPLS), the normalization method (Nor), the multiplicative scatter correction method (MSC), and the standard normal variant and standard normal variant transformation de-trending method (SNV-DT). Then, optimized characteristic variables were selected using the competitive adaptiive reweighted sampling method (CARS-SPA) and the backward interval partial least squares method (BiPLS). Based on that, a model for identification of edible oil was established using the support vector classification method (SVC). The results revealed that the SVC model established can accurately identify and classify eight different edible oil (soybean oil, peanut oil, rapeseed oil, tea seed oil, rice oil, corn oil, sunflower oil, and palm oil). The prediction accuracy for samples in calibration set and prediction set by the proposed model can be 100%, which is superior to that of conventional single-spectra model. The proposed model exhibits excellent generalization capability. Additionally, the study suggests that the Raman-NIR fusion shows improved efficiency in identification of edible oil and great potential for practical application.
会议论文集名称:
DEStech Transactions on Computer Science and Engineering
关键词:
Raman;Near infrared;Chemometrics;Edible oil;Iodine value
摘要:
Based on Raman and near infrared spectroscopy (NIR), the modeling of edible-oil iodine value was conducted following chemometrics and support vector machine regression (SVR). The Raman and NIR spectral data of 44 oil samples were collected and preprocessed. The preprocessed spectral data at characteristic wavelengths were extracted with CARS and iPLS methods. Parameter optimization was performed following the grid search algorithm (GS), for the SVR iodine value prediction models. The results show that all the models built could predict the iodine values to some extent. An NIR-MSC-iPLS-SVR model exhibited the greatest stability and its correlation coefficient R of the prediction set reached 97.69%. The results show that NIR has more advantages in the fast prediction of iodine values and can be utilized to manufacture portable spectral instruments for edible-oil iodine-value determination. Vegetable oils can afford human body unsaturated fatty acids, vitamins, and other nutrients [1]. Wherein, the unsaturated fatty acids can ensure the physiological function of cells, reduce cholesterol contents, and perfect blood circulation [2]. The unsaturated fatty acid content can be indicated by unsaturation degree and iodine value (iodine number). The iodine value refers to the grams of iodine consumed in the addition reaction of 100 g of oil. The higher the iodine value is, the greater the unsaturation degree is, and the higher the unsaturated fatty acid content is. The iodine value determination methods include Wijs method (national standard), high-performance liquid chromatography (HPLC) method, and so forth [3-4]. Wherein, the Wijs method is precise and commonly used. Nevertheless, because the sample and titrant solution are immiscible, at the end point of titration, the color change is too slow. And, the method is tedious and requires a large amount of hazardous organic solvents [5-6]. Compared with the Wijs method, the spectral method is fast, efficient, and free of sample pretreatment. Hence, it has been increasingly used in oil and food industries [7]. Yu Yanbo et al. [8] used near infrared spectroscopy (NIR) technology to build a prediction model for the prediction of contents of 4 fatty acids in vegetable oils. Maria A. Carmona et al. [9] identified oils and determined iodine values based on Raman spectroscopy.
会议名称:
Joint International Conference on Social Science and Environmental Science (SSES) / International Conference on Food Science and Engineering (ICFSE )
摘要:
This study aimed to analyze peanut oil adulteration based on chemometrics combined Raman and near-infrared (NIR) spectrum. The spectral data of 134 adulterated oil samples were collected by laser Raman and NIR spectrometer. The Raman spectra data and NIR were preprocessed respectively by different preprocessing approaches. The backward interval partial least squares (BiPLS) method was applied to extract the featured wavelengths. Based on the full spectrum and the characteristic wavelengths, adulteration quantity prediction models were established by the support vector machine regression (SVR) method. According to the analysis, the SVR model could predict the adulteration content in the peanut oil. Furthermore, the correlation coefficient R was greater than 0.97, and the mean square error (MSE) was smaller than 3.2E-4. The SVR model had advantageous properties such as strong generalization ability and good prediction accuracy. The results showed that Raman and NIR fusion analysis was effective in the quantitative analysis of the adulteration in peanut oil. Multi-spectral analysis for edible oil adulteration is an important field of study.
摘要:
Research on the use of Near-infrared Spectroscopy for rapid predict oil fatty acid content, proposed a method based on fatty acid content to identify the oil species. Collect five kinds of 133 parts of edible oil data samples by Near-infrared spectroscopy. The original spectroscopy were pretreated by using standard normal variable variation and De-trending(SNV-DT), and using support vector machine regression (SVR) build quantitative models of fatty acids, using Support Vector Classification (SVC) to establish the type of oil qualitative model. Results show, Using palmitic acid, oleic acid, linoleic three kinds of fatty acids it is feasible. Three kinds of fatty acid quantitative model prediction set correlation coefficients were 95.0876%,99.8592% and 98.5951%. Quantitative - Qualitative model prediction accuracy rate of 100% set. Studies shows that Near-infrared spectroscopy can quickly predict oil fatty acid content, and discriminating oil species. This research has a strong practical and popularization value. In order to develop a kind of rapid method which predict fatty acid content and use it to identify the oil species to provide technical support.