Multivariate Scatter Correction Research Articles

Non-destructive characterization and discrimination of vehicle bumpers fragments in forensic science using molecular spectral fusion analysis and chemometrics

A rapid and non-destructive approach for characterizing and discriminating trace evidences is one of the hotspots and focuses in forensic science. The vehicle bumpers fragments, as a trace evidence, are often found and extracted from the clothing of victims, the crashed vehicle, and the road at the scenes. Micro-laser Raman spectroscopy (MLRM) and attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR) has been used with principal component analysis (PCA), Multi-layer perceptron neural network (MLPNN), and Fisher discriminant analysis (FDA) to propose a method for characterizing and discriminating 160 vehicle bumpers fragments from 8 manufacturers. The spectra were pre-treated by automatic baseline correction, multivariate scatter correction, standard normal variate and Savitzky-Golay algorithm. PCA was used to reduce the dimension and extraction of feature variables, while MLPNN, FDA, the several supervised pattern recognition methods, were used as algorithms of constructing classifiers. The results displayed that the discrimination ability of fusion spectra model was stronger than that of single spectra dataset. In MLPNN model, the order of it from high to low was feature-level fusion (93.5%) > data-level fusion (90.3%) > MLRM (74.2%) > ATR-FTIR (64.5%). In FDA model, the order of it from high to low was feature-level fusion (100%) > data-level fusion (93.5%) > MLRM (77.4%) > ATR-FTIR (61.3%). Fisher discriminant analysis model based on feature-level fusion dataset was the more optimal and practical method for discriminating the vehicle bumpers fragments. “Audi”, “Hyundai”, and “Chery” samples, the similar kind from different manufacturers, were also distinguished exactly, which was satisfactory. The designed method represented a potentially rapid and non-destructive approach for differentiation trace evidences in forensic science.

Discrimination of Waxy Wheats Using Near-Infrared Hyperspectral Spectroscopy

Wheat (Triticum aestivum L.) carries three waxy loci (Wx-A1, Wx-B1, and Wx-D1) encoding granule-bound starch synthase (GBSS) which are related to amylose synthesis. The present study was to investigate the possibility of using near-infrared (NIR) hyperspectral spectroscopy to differentiate waxy wheat and three partial waxy wheats from wild-type wheat. Nearly 267 seeds from the same harvest year were used to obtain hyperspectral imaging maps in the near-infrared range (930–2548 nm), and then the data were analyzed based on chemometric methods. The first derivative, standard normal variable (SNV) transformation, and multivariate scattering correction (MSC) were used for spectral pretreatment. Support vector machine (SVM) and partial least square discriminant analysis (PLS-DA) and backpropagation neural network (BPNN) were applied to build discrimination models for wheat line classification. The results demonstrate that the highest classification accuracy of 98.51% for the prediction set was achieved by SVM model based on raw spectral data in full NIR region, and the classification accuracy of wild-type and partial waxy wheats all reached 100%. SVM models’ prediction accuracy (98.51%) is higher than PLS-DA and BPNN models’ (75.76% and 82.10%). The above results show that NIR hyperspectral spectroscopy combined SVM model was useful in identification of waxy and partial waxy wheats from wild-type wheat.

Hybrid variable selection strategy coupled with random forest (RF) for quantitative analysis of methanol in methanol-gasoline via Raman spectroscopy

With the trend of portable and miniaturization, Raman spectrometer requires more advanced analytical methods providing more rapid and accurate analysis performance for in-situ analysis. In this work, a hybrid variable selection method based on V-WSP and variable importance measurement (VIM) coupled with random forest (RF) was used to improve the quantitative analysis performance of portable laser Raman instruments for quantitative analysis of methanol content in methanol gasoline. First, five preprocessing methods were applied to reduce the infection information in the raw spectra, respectively. Based on the spectra data processed by multivariate scattering correction (MSC), V-WSP was employed to filter the infection or redundant information in Raman spectroscopy, and 579 variables were obtained when the correlation threshold is 0.9600. Then, the variables were further eliminated by VIM. Finally, 43 variables were obtained by the V-WSP-VIM method. In data processing, out of bag (OOB) error estimation and 10-flod cross validation (CV) were applied to optimize the parameters of preprocessing methods, V-WSP, VIM and RF model. The results fully demonstrated that compared with the RF model based on raw spectra, the RF model based on V-WSP-VIM method can achieve a better prediction performance for the quantitative analysis of methanol content in methanol-gasoline, with the coefficients of determination of cross-validation (R2CV) improving from 0.9100 to 0.9662, the root mean square error of cross-validation (RMSECV) reducing from 0.0572 to 0.0365%, the coefficients of determination of prediction set (R2P) improving from 0.9214 to 0.9407, the root mean square error of prediction set (RMSEP) reducing from 0.0420 to 0.0382%, the variables reducing from 1044 to 43 and the modeling time reducing from 72.94 to 6.41 s. The results indicates that V-WSP-VIM coupled with RF is an effective method to improve the performance of portable laser Raman spectrometer for quantitative analysis of methanol content in methanol gasoline.

A study of starch content detection and the visualization of fresh-cut potato based on hyperspectral imaging.

Fresh-cut potatoes are popular with consumers because of their healthiness, hygiene, and convenience. Currently, starch content is mainly detected using chemical methods, which are time-consuming and laborious. Moreover, these methods may cause some side effects in the human body. Therefore, suitable methods are required for the rapid and accurate detection of starch content. In this study, Zihuabai and Atlantic potatoes were used as experimental samples. The potatoes were sliced with stainless-steel blades, and images of these potatoes were obtained through hyperspectral imaging. The images were preprocessed using different methods. Competitive adaptive reweighed sampling (CARS) and the successive projection algorithm (SPA) were used to extract characteristic wavelengths. A partial least squares regression (PLSR) model was constructed to predict the starch content from the preprocessed full spectrum and the spectrum under the characteristic wavelength. The results indicate that the full spectrum model constructed through standard normal variable transformation (SNV) preprocessing had the best performance, with a correlation coefficient in the calibration set (Rc) value of 0.9020, a root mean square error of correction (RMSEC) of 2.06, and a residual prediction deviation (RPD) of 2.33. The characteristic wavelength-based multivariate scattering correction (MSC)-CARS-PLSR model exhibited better performance than the PLSR model constructed using the full spectrum, with an Rc value of 0.9276, RMSEC of 1.76, correlation coefficient in the prediction set (Rp) value of 0.9467, root mean square error of prediction of 1.63, and RPD of 2.95. The starch content in fresh-cut potatoes was visualized using the best model in combination with pseudocolor technology. The results indicate that hyperspectral imaging is effective for mapping the spatial distribution of starch content; thus, a solid theoretical basis is obtained for the grading and online monitoring of fresh-cut potato slices.

Research on the quantitative inversion model of heavy metals in soda saline land based on visible-near-infrared spectroscopy

With saline-alkali land in Zhenlai County, Baicheng city, Jilin Province, China as the research object, a quantitative inversion model for the heavy metal content of manganese (Mn), cobalt (Co) and iron (Fe) in saline-alkali soil and the visible-near-infrared spectroscopy data were established. First, Savitzky-Golay (SG) smoothing, multivariate scattering correction (MSC), continuum removal (CR), and combined transformation methods were performed on the original spectral data. By analysing the correlation between processed data and heavy metal content, the characteristic bands corresponding to different spectral transformations were extracted. Next, the ratio index (RI), difference index (DI) and normalized difference index (NDI) were constructed, and the spectral index was determined to have the most significant correlation with the combination of Mn-, Co- and Fe-content-sensitive bands and its corresponding Spearman rank correlation coefficient. Finally, a quantitative inversion model for the heavy metal content (Mn, Co and Fe) in soda saline-alkali land was established, and its accuracy was verified. The research results show that the optimal band selection principle for the quantitative inversion model of heavy metals in soda saline-alkali land is as follows: Mn selected 219 groups of sensitive bands with correlation coefficient r > 0.70, Co selected 1377 groups of sensitive bands with correlation coefficient r > 0.80, and Fe selected 104 groups of sensitive bands with correlation coefficient r > 0.80. Based on these selection principles, the random forest algorithm was used to perform an inversion for the Mn, Co and Fe contents with the best results, the goodness of fit (R2) values between predicted and measured values were 0.76, 0.92 and 0.91, and the mean relative accuracy was 91.75%, 92.45% and 93.90%, respectively. This method has improved prediction accuracy of the Mn, Co and Fe contents in saline-alkali land.

On the rapid and non-destructive approach for barbiturates, benzodiazepines, and phenothiazines determination and differentiation using spectral combination analysis and chemometric methods

The fast determination and differentiation of substance samples (tablets or powders) of medical drugs with high addiction potential by police or customs is an important forensic task. The sedative hypnotic drugs, due to their accessibility and low price, not only make some people from trying to use to becoming depended and addicted to them gradually, but also let some people be addicted to certain types of sedative hypnotic drugs due to the prolonged use. The attenuated total reflection (ATR)-Fourier transform infrared spectroscopy (FTIR) has been adopted with principal component analysis (KNN), Fisher discriminant analysis (FDA), and K nearest neighbor analysis (KNN) to develop a method for differentiating the target sedative hypnotic drugs of abuse. Automatic baseline correction, multivariate scatter correction, standard normal variate and Savitzky-Golay algorithm and the ATR correction were performed as pre-processing methods. PCA was used to reduce the infrared spectroscopy multi-dimensionality and extraction of feature variables. While FDA and KNN, the several supervised pattern recognition methods, were used as algorithms of constructing classifiers. The results showed that accuracy of combination spectral models was significantly higher than that of single models. The distinguishing ability of FDA model was superior to that of KNN model. First-second derivative model was better than that of other models in differentiating samples. All bands region model was more worthwhile than functional group or fingerprint region models. The optimal differentiation model was FDA model that based on all band region data of first-second derivative spectral combination. 13 different types of sedative hypnotic drugs were all differentiated exactly. The same kind of them from different places of origin were also distinguished accurately. The designed method represented a potentially rapid and non-destructive approach for determination and differentiation sedative hypnotic drugs.

Acquisition and analysis of hyperspectral data for surface contamination level of insulating materials

Severe contamination increases the risk of creeping discharge even flashover of the insulating materials. In pursuit of a reliable remote monitoring method to prevent unexpected flashover of contaminated insulating materials, this paper attempts to use hyperspectral imaging technology (HSI) with a targeted bands-based model to evaluate the contamination level. To validate such method, artificial samples with different contamination levels were employed as objects. By using Canny operator, pollution areas were selected as regions of interest. Then, multivariate scattering correction (MSC) was adopted in spectral data preprocessing and its effect was evaluated by principal component analysis (PCA) in terms of the data separability. By means of successive projection algorithm (SPA), six targeted bands were finally determined. In pursuit of a high discriminability of image expression, linear discriminate analysis (LDA) was selected and implemented on the targeted bands-based results. It demonstrated that MSC-SPA-LDA model is of a high discrimination of contamination level as well as a high calculation efficiency, which is significative on the rapid remote diagnosis based on HSI.

NIR Hyperspectral Imaging Technology Combined with Multivariate Methods to Identify Shrimp Freshness

In this study, a hyperspectral imaging system of 866.4–1701.0 nm, combined with a variety of spectral processing methods were adopted to identify shrimp freshness. To gain the optimal model combination, three preprocessing methods (Savitzky-Golay first derivative (SG1), multivariate scatter correction (MSC), and standard normal variate (SNV)), three characteristic wavelength extraction algorithms (random frog algorithm (RFA), uninformative variables elimination (UVE), and competitive adaptive reweighted sampling (CARS)), and four discriminant models (partial least squares discrimination analysis (PLS-DA), least squares support vector machine (LSSVM), random forest (RF), and extreme learning machine (ELM)) were employed for experimental study. First of all, due to the full wavelength modeling analysis, three preprocessing methods were utilized to preprocess the original spectral data. The analysis showed that the spectral data processed by the SNV method had the best performance among the four discriminant models. Secondly, due to the characteristic wavelength modeling analysis, three characteristic wavelength extraction algorithms were utilized to extract the characteristic wavelength of the SNV-processed spectral data. It was found that the CARS algorithm achieved the best performance among the three characteristic wavelength extraction algorithms, and the combining adoption of the ELM model and different characteristic wavelength extraction algorithms obtained the best results. Therefore, the model based on SNV-CARS-ELM obtained the best performance and was elected as the optimal model. Lastly, for accurately and explicitly displaying the refrigeration days of shrimps, the original hyperspectral images of shrimps were substituted into the SNV-CARS-ELM model, thus obtaining the general classification accuracy of 97.92%, and the object-wise method was used to visualize the classification results. As a result, the method proposed in this study can effectively detect the freshness of shrimps.

Forensic Classification of Pigments by Attenuated Total Reflectance – Fourier Transform Infrared Spectroscopy and Chemometrics

The rapid, nondestructive, and accurate classification of pigments in forensic science is important and indispensable. Here, a method for distinguishing different brands and types of pigments was developed by attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) with chemometrics. A total of 48 pigment samples were collected, and the corresponding infrared spectra were obtained. Baseline correction, multivariate scatter correction, standard normal variable analysis and Savitzky-Golay smoothing were used to preprocess the infrared spectra. Principal component analysis (PCA), factor analysis (FA), Laplacian eigenmaps (LE) and linear discriminant analysis (LDA) were used to extract the characteristic variables of the spectra for the samples. The results were classified by Bayesian discriminant analysis (BDA) and the K-nearest neighbor (KNN) method. The results show that BDA provided a more efficient and accurate model than KNN and the overall classification accuracy was almost 100.0%. Additionally, the classification model was more accurate after extracting the characteristic variables than with the direct use of BDA or KNN. The classification accuracy of gouache and acrylic pigments was 100.0% based on BDA and characteristic variables. The classification accuracy of the BDA and PCA model was 97.2% for two types of gouache pigments and two brands of Picasso gouache pigments. The results indicate that the combination of ATR-FTIR and BDA with a dimensionality reduction method is a potential tool for the classification of different brands and types of pigments.

Classification of heroin, methamphetamine, ketamine and their additives by attenuated total reflection-Fourier transform infrared spectroscopy and chemometrics.

Drug crime is a prominent issue of concern from pole to pole. In order to seek higher profits, drug gangs often add diluents and adulterants to the drugs to disperse drug products Analysis of these additives would be greatly conducive to determine the origin of drug products for law enforcement departments. A method using attenuated total reflectance-Fourier transform infrared spectroscopy and chemometrics methods to classify the heroin hydrochloride, methamphetamine hydrochloride, ketamine hydrochloride and their five additives (caffeine, phenacetin, starch, glucose, and sucrose), was developed. The Baseline correction, multivariate scatter correction, standard normal variate and Savitzky-Golay algorithm were adopted to pre-process the spectral data. Several supervised pattern recognition methods including decision tree, Bayes discriminant analysis, and support vector machine were considered as algorithms of constructing classifiers. The results reveal that, repetitive and interfering data in original spectrum data could be eliminated by principal component analysis and factor analysis. F-measure, as a comprehensive evaluation index of precision rate and recall rate, was more objective than precision rate and recall rate to reflect the ability of model to distinguish samples. It should be used as one of the indicators to evaluate the model. The CHAID classification tree could be identified as priorities in the decision tree model, while the linear kernel could be considered as the optimal kernel in the support vector machine model. The classification ability of three hydrochloride mixtures based on Bayes discriminant analysis was better than that of another models. Bayes discriminant analysis model was the more useful and practical method for classifying the target drugs of abuse than that of decision trees and support vector machine. The designed approach represents a potentially simple, non-destructive, and rapid method of classifying hydrochloride mixtures.

Detection of Three-component Adulteration in Camellia Oleifera Abel Seed Oil by Near Infrared Spectroscopy

In this paper, A combination of near-infrared microscopy and chemometrics were employed to identify the adulteration in Camellia oleifera Abel seed oil. The qualitative and quantitative detection methods for three-component adulteration in Camellia oleifera Abel seed oil was studied, and the models for the authenticity identification and content determination of Camellia oleifera Abel seed oil were established. Multivariate scattering correction (MSC), standard normal variate (SNV) and combination of standard normal variate and de-trend (SNV-DT) were separately used for the preprocessing of original spectra. Backward interval partial least squares (BiPLS) method was used to optimize feature variables, and partial least squares (PLS) method was employed to extract the spectrum from the characteristic wavelength. The parameters were optimized through comparative interactive verification (CV) and genetic algorithm (GA). With the establishment of qualitative and quantitative models in the full wavelength and characteristic wavelength ranges, the qualitative and quantitative identification paths were optimized. The results show that the identification and quantitative detection of Camellia oleifera Abel seed oil adulterated with three dopants could be realized by NIRS. The accuracy of as-established MSC-BiPLS-GA-SVC model was up to 100%. The correlation coefficient R of the MSC-BiPLS-GA-SVC model reached 99.7254%, and the MSE value was 0.0234.

Development of Visible/Near-Infrared Hyperspectral Imaging for the Prediction of Total Arsenic Concentration in Soil

Soil total arsenic (TAs) contamination caused by human activities—such as mining, smelting, and agriculture—is a problem of global concern. Visible/near-infrared (VNIR), X-ray fluorescence spectroscopy (XRF), and laser-induced breakdown spectroscopy (LIBS) do not need too much sample preparation and utilization of chemicals to evaluate total arsenic (TAs) concentration in soil. VNIR with hyperspectral imaging has the potential to predict TAs concentration in soil. In this study, 59 soil samples were collected from the Daye City mining area of China, and hyperspectral imaging of the soil samples was undertaken using a visible/near-infrared hyperspectral imaging system (wavelength range 470–900 nm). Spectral preprocessing included standard normal variate (SNV) transformation, multivariate scatter correction (MSC), first derivative (FD) preprocessing, and second derivative (SD) preprocessing. Characteristic bands were then identified based on Spearman’s rank correlation coefficients. Four regression models were used for the modeling prediction: partial least squares regression (PLSR) (R2 = 0.71, RMSE = 0.48), support vector machine regression (SVMR) (R2 = 0.78, RMSE = 0.42), random forest (RF) (R2 = 0.78, RMSE = 0.42), and extremely randomized trees regression (ETR) (R2 = 0.81, RMSE = 0.38). The prediction results were compared with the results of atomic fluorescence spectrometry methods. In the prediction results of the models, the accuracy of ETR using FD preprocessing was the highest. The results confirmed that hyperspectral imaging combined with Spearman’s rank correlation with machine learning models can be used to estimate soil TAs content.

Integration of Partial Least Squares Regression and Hyperspectral Data Processing for the Nondestructive Detection of the Scaling Rate of Carp (Cyprinus carpio).

The scaling rate of carp is one of the most important factors restricting the automation and intelligence level of carp processing. In order to solve the shortcomings of the commonly-used manual detection, this paper aimed to study the potential of hyperspectral technology (400–1024.7 nm) in detecting the scaling rate of carp. The whole fish body was divided into three regions (belly, back, and tail) for analysis because spectral responses are different for different regions. Different preprocessing methods, including Savitzky–Golay (SG), first derivative (FD), multivariate scattering correction (MSC), and standard normal variate (SNV) were applied for spectrum pretreatment. Then, the successive projections algorithm (SPA), regression coefficient (RC), and two-dimensional correlation spectroscopy (2D-COS) were applied for selecting characteristic wavelengths (CWs), respectively. The partial least square regression (PLSR) models for scaling rate detection using full wavelengths (FWs) and CWs were established. According to the modeling results, FD-RC-PLSR, SNV-SPA-PLSR, and SNV-RC-PLSR were determined to be the optimal models for predicting the scaling rate in the back (the coefficient of determination in calibration set (RC2) = 96.23%, the coefficient of determination in prediction set (RP2) = 95.55%, root mean square error by calibration (RMSEC) = 6.20%, the root mean square error by prediction (RMSEP)= 7.54%, and the relative percent deviation (RPD) = 3.98), belly (RC2 = 93.44%, RP2 = 90.81%, RMSEC = 8.05%, RMSEP = 9.13%, and RPD = 3.07) and tail (RC2 = 95.34%, RP2 = 93.71%, RMSEC = 6.66%, RMSEP = 8.37%, and RPD = 3.42) regions, respectively. It can be seen that PLSR integrated with specific pretreatment and dimension reduction methods had great potential for scaling rate detection in different carp regions. These results confirmed the possibility of using hyperspectral technology in nondestructive and convenient detection of the scaling rate of carp.

Characterization of Heroin and Its Additives by Attenuated Total Reflection (ATR) – Fourier Transform Infrared Spectroscopy (FTIR) and Multivariate Analysis

A method using attenuated total reflectance (ATR) Fourier transform infrared (FTIR) spectroscopy and multivariate analysis is reported to classify heroin hydrochloride and five common additives: caffeine, phenacetin, starch, glucose, and sucrose. Baseline correction, multivariate scatter correction, standard normal variables and Savitzky-Golay algorithm using smoothing with a polynomial order of three and a window size of seven points were adopted to preprocess the spectral data. Several supervised pattern recognition methods including radial basis function neural network (RBFNN), multilayer perceptron neural network (MLPNN), and linear fitting analysis (LFA) were used as algorithms to construct the classifiers. The use of a characteristic spectrum had lower complexity, required reduced modeling time, and was able to achieve the goal of rapid classification. Mixtures of heroin hydrochloride with caffeine, heroin hydrochloride and phenacetin, heroin hydrochloride with sucrose, heroin hydrochloride and starch, and heroin hydrochloride with glucose were distinguished with accuracy values of 100%, 100%, 88.89%, 77.78% and 66.67%, respectively. The fitting ability of the quadratic polynomial function was superior than using a linear model. A linear model was shown to be optimal when the additive was phenacetin, although the quadratic function was more superior when the additive was glucose. The developed method represents a simple, nondestructive, and rapid approach to classify mixtures of heroin hydrochloride with its common additives, caffeine, glucose, phenacetin, starch and sucrose.

Hyperspectral imaging for accurate determination of rice variety using a deep learning network with multi-feature fusion.

The phenomena of rice adulteration and shoddy rice arise continuously in high-quality rice and reduce the interests of producers, consumers and traders. Hyperspectral imaging (HSI) was conducted to determine rice variety using a deep learning network with multiple features, namely, spectroscopy, texture and morphology. HSI images of 10 representative high-quality rice varieties in China were measured. Spectroscopy and morphology were extracted from HSI images and binary images in region of interest, respectively. And texture was obtained from the monochromatic images of characteristic wavelengths which were highly correlated with rice varieties. A deep learning network, namely principal component analysis network (PCANet), was adopted with these features to develop classification models for determining rice variety, and machine learning methods as K-nearest neighbour and random forest were used to compare with PCANet. Meanwhile, multivariate scatter correction, standard normal variate, Savitzky-Golay smoothing and Savitzky-Golay's first-order were applied to eliminate spectral interference, and principal component analysis (PCA) was performed to obtain the main information of high-dimensional features. Multi-feature fusion improved recognition accuracy, and PCANet demonstrated considerable advantage in classification performance. The best result was achieved by PCANet with PCA-processed spectroscopic and texture features with correct classification rates of 98.66% and 98.57% for the training and prediction sets, respectively. In summary, the proposed method provides an accurate identification of rice variety and can be easily extended to the classification, attribution and grading of other agricultural products.

Hyperspectral imaging technology combined with multivariate data analysis to identify heat-damaged rice seeds

Seeds in the drying process due to improper control of process and temperature often cause heat damage to affect seed vigor. How to quickly and nondestructively identify seeds that are heat-damaged plays a key role in agricultural production. In this study, the electric heating constant temperature blast drying oven was used to simulate the drying process of rice seeds and heated for different time (Untreated, 1 h/60 °C, 3 h/60 °C, 5 h/60 °C, 7 h/60 °C, 9 h/60 °C, 11 h/60 °C). The hyperspectral images of rice seeds with different degrees of heat damage were obtained by using a hyperspectral imaging system of 866.4–1701.0 nm. Three preprocessing methods (Savitzky-Golay first derivative, standard normal variate, and multivariate scatter correction) were used to preprocess the original spectral data, three feature extraction algorithms (second derivative, successive projections algorithm, and neighborhood component analysis) were used to extract the feature wavelengths, and three classifier models (k-nearest neighbor, support vector machine, and naive Bayes) were used for modeling analysis. After multivariate data analysis, the multivariate scatter correction-neighborhood component analysis-naive Bayes model performed best and was selected as the best model. Finally, the hyperspectral images of the verification set were visualized based on the object-wise method to show the intuitive classification effect. The results show that the hyperspectral imaging technology is an effective tool for quickly identifying and visualizing rice seeds with different degrees of heat damage.

Near-infrared-based Identification of Walnut Oil Authenticity

Near-infrared spectroscopy (NIR) was employed to determine the quality of walnut oil. Multivariate scattering correction (MSC) and combination of standard normal variate and de-trend (SNV-DT) were separately used for preprocessing. Successive projection algorithm (SPA) and combination of competitive adaptive reweighted sampling algorithm and partial least squares method (CARS-PLS) were separately used for extraction of characteristic wavelengths. Particle swarm optimization (PSO) and genetic algorithm (GA) methods were separately used for parameter optimization. Eventually, quantitative prediction models for impurities contents in walnut oil were established by support vector machine (SVM). After analysis of the prediction results derived from these models, an optimal candidate was obtained.

Chlorophyll content for millet leaf using hyperspectral imaging and an attention-convolutional neural network

ABSTRACT: Chlorophyll is a major factor affecting photosynthesis; and consequently, crop growth and yield. In this study, we devised a chlorophyll-content detection model for millet leaves in different stages of growth based on hyperspectral data. The hyperspectral images of millet leaves were obtained under a wavelength range of 380-1000 nm using a hyperspectral imager. Threshold segmentation was performed with near-infrared (NIR) reflectance and normalized difference vegetation index (NDVI) to intelligently acquire the regions of interest (ROI). Furthermore, raw spectral data were preprocessed using multivariate scatter correction (MSC). A correlation coefficient-successive projections algorithm (CC-SPA) was used to extract the characteristic wavelengths, and the characteristic parameters were extracted based on the spectral and image information. A partial least squares regression (PLSR) prediction model was established based on the single characteristic parameter and multi-characteristic parameter fusion. The determination coefficient (Rv 2) and the root-mean-square error (RMSEv) of the validation set for the multi-characteristic parameter fusion model were reported to be 0.813 and 1.766, respectively, which are higher than those obtained by the single characteristic parameter model. Based on the multi-characteristic parameter fusion, an attention-convolutional neural network (attention-CNN) (Rv 2 = 0.839, RMSEv = 1.451, RPD = 2.355) was established, which is more effective than the PLSR (Rv 2 = 0.813, RMSEv = 1.766, RPD = 2.167) and least squares support vector machine (LS-SVM) models (Rv 2 = 0.806, RMSEv = 1.576, RPD = 2.061). These results indicated that the combination of hyperspectral imaging and attention-CNN is beneficial to the application of nutrient element monitoring of crops.

Hyperspectral imaging technology combined with deep forest model to identify frost-damaged rice seeds.

In recent years, deep learning models have been widely used in the field of hyperspectral imaging. However, the training of deep learning models requires not only a large number of samples, but also the need to set too many hyper-parameters, which is time consuming and laborious for researchers. This study used hyperspectral imaging technology combined with a deep learning model suitable for small-scale sample data sets, deep forests (DF) model, to classify rice seeds with different degrees of frost damage. During the period, three spectral preprocessing methods (Savitzky-Golay first derivative (SG1), standard normal variate (SNV), and multivariate scatter correction (MSC)) were used to process the original spectral data, and three feature extraction algorithms (principal component analysis (PCA), successive projections algorithm (SPA), and neighborhood component analysis (NCA)) were used to extract the characteristic wavelengths. Moreover, DF model and three traditional machine learning models (decision tree (DT), k-nearest neighbor (KNN), and support vector machine (SVM)) were built based on different numbers of sample sets. After multivariate data analysis, it showed that the pretreatment effect of MSC was the most excellent, and the characteristic wavelength extracted by NCA algorithm was the most useful. In addition, the performance of DF model was better than these three traditional classifier models, and it still performed well in small-scale sample set data. Therefore, DF model was chosen as the best classification model. The results of this study show that the DF model maintains good classification performance in small-scale sample set data, and it has a good application prospect in hyperspectral imaging technology.

Rapid and Nondestructive Forensic Identification of Tire Particles by Attenuated Total Reflectance – Fourier Transform Infrared Spectroscopy and Chemometrics

It is of great significance to realize the rapid, nondestructive and accurate identification of tire rubber particles in forensic science. However, there are no relevant reports. This study collected and tested infrared spectroscopy data of 240 samples from 15 brands. Baseline correction, multivariate scatter correction, standard normal variables and Savitzky-Golay smoothing were adopted to preprocess the infrared spectra. The comparison of the fingerprint region and full infrared spectrum was explored. The identification models were established by principal component analysis and discriminant analysis. The results showed that there is no significant discrimination in the peak position and shape; instead the peak intensity and relative height were slightly different. The cumulative variance contribution rate of principal components 1 through 12 reached 98.3690% demonstrating 98.3690% of the details in the original measurements. The 39-dimensional principal components of infrared fingerprint region accurately differentiated sample brands and employed computational complexity. Styrene butadiene samples from six brands were accurately identified based on the three discriminant functions Za1, Zb1, and Zc1. Butadiene samples from five brands were differentiated using discriminant functions Za2, Zb2, and Zc2. Isoprene samples from brands were separated using the discriminant functions Za3, Zb3, and Zc3. All samples were precisely identified by this rapid and nondestructive protocol. The results of this study demonstrate the potential of attenuated total reflectance-Fourier transform infrared spectroscopy in combination with principal component analysis and discriminant analysis as a new method for the identification of tire rubber particles.