Calibration Correlation Research Articles

Evaluation of Carbohydrate Concentrations in Phalaenopsis Using Near-infrared Spectroscopy

Carbohydrate concentrations are important indicators of the internal quality of Phalaenopsis. In this study, near-infrared (NIR) spectroscopy was used for quantitative analyses of fructose, glucose, sucrose, and starch in Phalaenopsis plants. Both modified partial least-squares regression (MPLSR) and stepwise multiple linear regression (SMLR) methods were used for spectral analysis of 302 Phalaenopsis samples in the full visible NIR wavelength range (400–2498 nm). Calibration models built by MPLSR were better than those built by SMLR. For fructose, the smoothed first derivative MPLSR model provided the best results, with a correlation coefficient of calibration (Rc) of 0.96, standard error of calibration (SEC) of 0.22% dry weight (DW), standard error of validation (SEV) of 0.28% DW, and bias of -0.01% DW. For glucose, the MPLSR model based on the smoothed first derivative spectra was the best (Rc = 0.96; SEC = 0.26% DW; SEV = 0.32% DW; and bias = 0.01% DW). The best MPLSR model of sucrose was developed using the smoothed first derivative spectra (Rc = 0.96; SEC = 0.24% DW; SEV = 0.31% DW; bias = -0.03% DW). Regarding starch, the smoothed first derivative MPLSR model showed the best effects (Rc = 0.91; SEC = 0.47% DW; SEV = 0.56% DW; bias = -0.02% DW). Both the MPLSR and SMLR models showed satisfactory predictability, indicating that NIR has the potential to be adopted as an effective method of rapid and accurate inspection of the carbohydrate concentrations of Phalaenopsis plants. This technique could contribute substantially to quality management of Phalaenopsis.

Validation of a Multiresidue Analysis Method for 379 Pesticides in Human Serum Using Liquid Chromatography-Tandem Mass Spectrometry.

A screening method for simultaneous analysis of 379 pesticides in human serum was developed using liquid chromatography-tandem mass spectrometry (LC-MS/MS). Electrospray ionization with positive/negative switching mode of LC-MS/MS was adopted, and scheduled multiple reaction monitoring for each target compound was established. The limit of quantitation was 10 ng/mL for 94.5% of the total pesticides, and the correlation coefficients of calibration were ≥0.990 for 93.9% of the pesticides. For the sample preparation, scaled-down QuEChERS were used. Serum (100 μL) was extracted with acetonitrile (400 μL), partitioned with magnesium sulfate (40 mg) and sodium chloride (10 mg), and the upper layer was used for analysis without further cleanup steps. For the accuracy and precision tests, most of the pesticides showed excellent results in intra- and interday conditions. In the recovery tests at 10, 50, and 250 ng/mL, 85.8-91.8% of all target compounds satisfied the recovery range of 70-120% (relative standard deviation ≤20%).

Capacità predittiva della tecnologia FT-NIRs della composizione chimica di prodotti stagionati di suino

Conventional chemical analyses of meat cured products are time-consuming, expensive and destructive. Advantages of NIR spectroscopy are its speed, simplicity, low-cost and the possibility to determine a large number of different parameters simultaneously in a large number of samples. The aim of this study is to assess the ability of FT-NIRs to predict chemical composition of seasoned pig products. One hundred and two seasoned products were sampled (43 Cuore di spalla, 26 Toscano Ham and 33 Capocollo) and the following chemical components were determined: protein, intramuscular fat, ashes and fatty acid composition. NIR spectra were collected using a Thermo-Fisher Antaris II instrument. Partial least squares (PLS) regression was applied in the calibration and the validation models; the models were fully cross-validated using the “leave-one-out” method. Calibration and cross validation models were developed both for each product separately and grouping all the samples. Calibration correlation coefficients show satisfying values (minimum R2=0.73), while cross-validation correlation coefficients, despite being generally acceptable, show lower values (minimum R2=0.42). Best R2 was found for fat content (cross-validation R2= 0.95). Results, even if obtained on a reduced sample, show how FT-NIRs could be used in routine analyses of pig seasoned products.

Geographical origin discrimination and polysaccharides quantitative analysis of Radix codonopsis with micro near-infrared spectrometer engine

At present, Tradition Chinese Medicine (TCM) industry in China is in the stage from the empirical development to industrial production. Near infrared (NIR) spectroscopy has been widely used in the quality control of TCM’s modernization with its characteristics including rapidness, nondestruction, simplicity, economy, and so on. In this study, as one type of a portable micro NIR spectrometer, Micro NIR 1700 was used to establish the qualitative models for identification of geographical region and authenticity of Radix codonopsis based on discriminant analysis (DA) method. Both of the DA models had better predictive ability with 100% accuracy. In addition, a method for rapid quantitative analysis of polysaccharide in Radix codonopsis was also developed based on Micro NIR 1700 spectrometer with partial least-squares (PLS) algorithm. In the PLS calibration model, the NIR spectra of samples were pretreated with different preprocessing methods and the spectral region was selected with different variable selection methods as well. The performance of the final PLS model was evaluated according to correlation coefficient of calibration ([Formula: see text]), correlation coefficient of prediction ([Formula: see text]), root mean squared error of cross validation (RMSECV), and root mean squared of prediction (RMSEP). The values of [Formula: see text], [Formula: see text], RMSECV, and RMSEP were 0.9775, 0.9602, 2.496, and 2.734[Formula: see text]g/mL, respectively. This work demonstrated that micro infrared spectrometer could be more convenient and rapid for quality control of Radix codonopsis, and the presented models would be a useful reference for quality control of other similar raw materials of TCM.

Quantitative evaluation of pork marbling score along Longissimus thoracis using NIR images of rib end

This study investigated the potential of using the image of the rib end to assess pork marbling scores along Longissimus thoracis. Previous studies have shown the success of hyperspectral imaging in marbling score assessment of individual pork chops. In this research, hyperspectral imaging methods were employed to study the possibility of pork marbling assessment along longissimus thoracis using the NIR image of the rib end. Pattern analysis algorithms were used to extract efficient features from the rib end of the pork loin. Multiple linear regression (MLR) and leave-one-out cross validation were applied to build prediction models using the extracted features. Experiments showed that the prediction models performed well with correlation coefficients of calibration (Rc) ≥ 0.94, correlation coefficients of cross validation (Rcv) ≥ 0.88, correlation coefficients of prediction (Rp) ≥ 0.89. The promising results indicated the potential of using NIR images of rib ends for the assessment of marbling score along longissimus thoracis.

Determination and Visualization of Peimine and Peiminine Content in Fritillaria thunbergii Bulbi Treated by Sulfur Fumigation Using Hyperspectral Imaging with Chemometrics.

Rapid, non-destructive, and accurate quantitative determination of the effective components in traditional Chinese medicine (TCM) is required by industries, planters, and regulators. In this study, near-infrared hyperspectral imaging was applied for determining the peimine and peiminine content in Fritillaria thunbergii bulbi under sulfur fumigation. Spectral data were extracted from the hyperspectral images. High-performance liquid chromatography (HPLC) was conducted to determine the reference peimine and peiminine content. The successive projection algorithm (SPA), weighted regression coefficient (Bw), competitive adaptive reweighted sampling (CARS), and random frog (RF) were used to select optimal wavelengths, while the partial least squares (PLS), least-square support vector machine (LS–SVM) and extreme learning machine (ELM) were used to build regression models. Regression models using the full spectra and optimal wavelengths obtained satisfactory results with the correlation coefficient of calibration (rc), cross-validation (rcv) and prediction (rp) of most models being over 0.8. Prediction maps of peimine and peiminine content in Fritillaria thunbergii bulbi were formed by applying regression models to the hyperspectral images. The overall results indicated that hyperspectral imaging combined with regression models and optimal wavelength selection methods were effective in determining peimine and peiminine content in Fritillaria thunbergii bulbi, which will help in the development of an online detection system for real-world quality control of Fritillaria thunbergii bulbi under sulfur fumigation.

Parameter optimization in soluble solid content prediction of entire bunches of grape based on near infrared spectroscopic technique

The object of this research was to evaluate grape’s soluble solid content (SSC) nondestructively based on near infrared spectroscopic technique using a detection probe designed in house and focus on optimization of three detection parameters that were light power (P), outer diameter of the probe (D), distance between the light source to the probe (L). Statistical models between diffuse transmittance spectra with grape’s SSC were developed using partial least square (PLS) regression. Orthogonal experiment was applied to choose optimal parameters. Comprehensive performance of different models under each combination of factor levels was assessed in terms of correlation coefficient of calibration (r c ), root mean square error of prediction (RMSEP) and differential values of RMSEP and root mean square error of calibration (Δ). The PLS model obtained the best results with the r c of 0.83, the RMSEP of 0.76 °Brix and the Δ of 0.84 °Brix on condition that P, D and L were 70 W, 70 and 85 mm, respectively. The results in the study show that SSC prediction of grape using the testing probe is feasible and parameter L has more impact on the performance of models than others. Better models can be obtained through reasonable parameters combination. This research can provide reference for the nondestructive detection of the entire bunches of grapes.

Rapidly Simultaneous Determination of Six Effective Components in Cistanche tubulosa by Near Infrared Spectroscopy.

Quantitative determination of multiple effective components in a given plant usually requires a very large amount of authentic natural products. In this study, we proposed a rapid and non-destructive method for the simultaneous determination of echinacoside, verbascoside, mannitol, sucrose, glucose and fructose in Cistanche tubulosa by near infrared spectroscopy (NIRS). Near infrared diffuse reflectance spectroscopy (DRS) and high performance liquid chromatography (HPLC) were conducted on 116 batches of C. tubulosa samples. The DRS data were processed using standard normal variety (SNV) and multiplicative scatter correction (MSC) methods. Partial least squares regression (PLSR) was utilized to build calibration models for components-of-interest in C. tubulosa. All models were then assessed by calculating the root mean square error of calibration (RMSEC), correlation coefficient of calibration (r). The r values of all six calibration models were determined to be greater than 0.94, suggesting each model is reliable. Therefore, the quantitative NIR models reported in this study can be qualified to accurately quantify the contents of six medicinal components in C. tubulosa.

A screen-printed voltammetric electronic tongue for the analysis of complex mixtures of metal ions

A voltammetric electronic tongue was constituted by four screen-printed modified electrodes: a carbon nanofiber modified electrode, an ex-situ antimony film electrode prepared from carbon nanofiber modified electrode, and two carbon nanofiber electrodes chemically modified with Cys and GSH. The tongue was successfully applied to the analysis of a complex mixture of metal ions (4 analytes and 2 interferences) by differential pulse anodic stripping voltammetry. Each sensor was firstly studied for the determination of each metal separately confirming that all electrodes showed differentiated response for the metals. The obtained voltammetric signals provided by the sensor array were processed by Partial Least Squares regression (PLS) to resolve the overlapped nature of the obtained multimetal stripping measurements. This PLS model was built considering a hierarchical model in order to reduce the large amount of data. The method was applied to synthetic mixtures of Cd(II), Pb(II), Tl(I), and Bi(III) in the presence of Zn(II) and In(III) at the levels of μg L−1 and successfully validated with correlation coefficients of both calibration and prediction higher than 0.9 obtained from predicted vs. expected concentration graphs. Moreover, the simultaneous determination of Cd(II), Pb(II), Tl(I), and Bi(III) in the presence of Zn(II) and In(III) in a spiked tap water was also satisfactory achieved, providing comparable results to those obtained by ICP-MS.

Experimental design-artificial neural network-genetic algorithm optimization and computer-assisted design of celecoxib molecularly imprinted polymer/carbon nanotube sensor

A molecularly imprinted polymer (MIP) sensor was prepared applying electropolymerization of pyrrole monomer in the presence of celecoxib (CXB) as a template molecule on a pencil graphite electrode (PGE) which modified with functionalized multi-walled carbon nanotubes (MWCNTs). The fabrication of sensor and the measurement process were carried out by cyclic voltammetry (CV) and differential pulse voltammetry (DPV) method, respectively. The selection of functional monomer was done using computational method. The optimization of parameters was performed using Plackett–Burman design (PBD), central composite design (CCD), artificial neural network (ANN) and genetic algorithm (GA). First, a Plackett–Burman design as a screening method was applied to estimate positive or negative influences of the factors on the sensor response. In the next step of optimization process, significant factors affecting the performance of the MIP sensor were investigated using CCD. The data from CCD used to train the ANN using Bayesian regularization. The ANN was trained with mean square error (MSE) of 0.0336. The predicted model obtained from the trained ANN was introduced to GA as the fitness function to be optimized. GA optimized the fitness function using the defined constraints. Under the obtained optimal conditions a linear calibration graph in the range 5.0×10−3–20μM was obtained with a calibration correlation coefficient of 0.9985. The limit of detection (3sb/slope) was obtained 2.34×10−3μM. The real sample analysis was performed to determine CXB in pharmaceutical and human serum samples. The repeatability of method (n=3) were obtained 3.03%.

Modeling of process indices of Aloe vera gel extraction machine using artificial neural network

AbstractThe performance evaluation of Aloe vera gel extraction machine in terms of extraction efficiency, gel recovery, and extraction loss were evaluated at three level of roller speeds (11.40, 15.10, and 18.80 rpm) and roller clearance (4, 5, and 6 mm). The data obtained were used for development of artificial neural network (ANN) model. The roller speed, roller clearance, extraction efficiency, gel recovery, and extraction loss were used as the inputs and the gel extraction efficiency, gel recovery, and extraction loss were the output from the model. New feed forward neural network was used for the model building and trained using Bayesian regularization training algorithm. Optimum ANN architecture was obtained by trial and error method. The results showed that the correlation coefficient for calibration and validation was .9751 and .9982, respectively, the root mean square error for calibration and validations was .4746 and .1258, respectively, and the ANN efficiency for calibration and validation was .9507 and .9964, respectively, were recorded for ANN model 5‐10‐3. Results indicated that ANN can be used efficiently for the modeling of the process indices of the Aloe vera gel extraction machine.Practical applicationsUsing a set of experimental data from Aloe vera gel extraction machine artificial neural network (ANN) model was developed to evaluate and predict the performance of Aloe vera gel extraction machine in terms of extraction efficiency, gel recovery, and extraction loss. The evaluation of ANN model with test data proved that model has very good accuracy. Using this model one can search for optimum operation conditions of Aloe vera gel extraction machine to maximize the efficiency and gel recovery and minimize the extraction loss.

Quantitative Determination of Ferulic Acid Content in Chrysanthemum Morifolium cv . (Chuju) Continuous Cropping Soil Using Near Infrared Spectroscopy

Abstract A near infrared spectroscopy (NIRS) method was used for rapid quality evaluation of ferulic acid content in chrysanthemum morifolium cv . (Chuju) continuous cropping soil. Standard leverage, studentized residual and Mahalanobis distance were calculated to eliminate abnormal samples. After the initial near infrared spectrum was treated by two second derivative and Norris smoothing filter noise, 6000–4000 cm −1 wave number range and 7 factors were chosen for partial least squares (PLS) calibration model. The results showed that good correlations were presented between the calibration set/validation set and the HPLC values, and the calibration correlation coefficient ( R c ) and validation correlation coefficient ( R cv ) were 0.9914 and 0.9935. Root mean square error of calibration (RMSEC), root mean square error of prediction (RMSEP) and root mean square error of cross prediction (RMSECV) were 0.484, 0.539 and 0.615, respectively. This method was accurate, simple, rapid, nondestructive, and suitable for analysis of ferulic acid in continuous cropping soil.

Accuracy assessment of near-shore bathymetry information retrieved from Landsat-8 imagery

The improvement in the capabilities of Landsat-8 imagery to retrieve bathymetric information in shallow coastal waters was examined. Landsat-8 images have an additional band named coastal/aerosol, Band 1: 435–451 nm in comparison with former generation of Landsat imagery. The selected Landsat-8 operational land image (OLI) was of Chabahar Bay, located in the southern part of Iran (acquired on February 22, 2014 in calm weather and relatively low turbidity). Accurate and high resolution bathymetric data from the study area, produced by field surveys using a single beam echo-sounder, were selected for calibrating the models and validating the results. Three methods, including traditional linear and ratio transform techniques, as well as a novel proposed integrated method, were used to determine depth values. All possible combinations of the three bands [coastal/aerosol (CB), blue (B), and green (G)] have been considered (11 options) using the traditional linear and ratio transform techniques, together with five model options for the integrated method. The accuracy of each model was assessed by comparing the determined bathymetric information with field measured values. The standard error of the estimates, correlation coefficients (R 2 ) for both calibration and validation points, and root mean square errors (RMSE) were calculated for all cases. When compared with the ratio transform method, the method employing linear transformation with a combination of CB, B, and G bands yielded more accurate results (standard error = 1.712 m, R 2 calibration = 0.594, R 2 validation = 0.551, and RMSE =1.80 m). Adding the CB band to the ratio transform methodology also dramatically increased the accuracy of the estimated depths, whereas this increment was not statistically significant when using the linear transform methodology. The integrated transform method in form of Depth = b 0 + b 1 X CB + b 2 X B + b 5 ln(R CB )/ln(R G ) + b 6 ln(R B )/ln(R G ) yielded the highest accuracy (standard error = 1.634 m, R 2 calibration = 0.634, R 2 validation = 0.595, and RMSE = 1.71 m), where R i (i = CB, B, or G) refers to atmospherically corrected reflectance values in the i th band [X i = ln(R i -R deep water)].

FIRST DERIVATIVE PREDICTION OF RAW BROILER SHEAR FORCE USING VISIBLE SHORT WAVE NEAR INFRARED SPECTROSCOPY

A non-destructive,fast, reliable and low cost technique which is Near-Infrared Spectroscopy (NIRS) is required to replace conventional destructive texture analyser in shear force measurement. The combination of visible and shortwave near infrared (VIS-SWNIR) spectrometer and principal component regression (PCR) to assess the quality attribute of raw broiler meat texture (shear force value (kg)) was investigated. Wavelength region of visible and shortwave 662-1005 nm was selected for prediction after pre-processing. Absorbance spectra was pre-processed using the optimal Savitzky-Golay smoothing mode with 1st order derivative, 2nd degree polynomial and 31 filter points to remove the baseline shift effect. Potential outliers were identified through externally studentised residual approach. The PCR model were trained with 90 samples in calibration and validated with 44 samples in prediction datasets. From the PCR analysis, correlation coefficient of calibration (RC), the root mean square calibration (RMSEC), correlation coefficient of prediction (RP) and the root mean square prediction (RMSEP) of visible and shortwave (662-1005 nm) with 4 principal components were 0.4645,0.0898, 0.4231 and 0.0945. The predicted results can be improved by applying the 2nd order derivative and the non-linear model.

Assessment of intramuscular fat content of pork using NIR hyperspectral images of rib end

A rapid and effective near infrared (NIR; 900–1700 nm) hyperspectral imaging method to determine intramuscular fat (IMF) content of pork at the last 6 ribs of the round rib end was developed and tested. Pattern analysis techniques including Gabor filter, grey-level co-occurrence matrix (GLCM), and wide line detector (WLD) were applied to process hyperspectral images into data from which spectral and image features, including texture-spectral, texture and line features were extracted. Among all spectral and image features of the 6 ribs' images, the first derivative of Gabor filtered mean spectra showed the strongest correlation with IMF content, and was therefore selected as the optimal feature for prediction of IMF content at any given rib number. Multiple linear regression (MLR) was exploited to build prediction models between IMF content at rib end and at 6 different rib locations. Leave-one-out cross validation was used to test the robustness of established models. The correlation coefficients of calibration, cross validation and full validation (Rc, Rcv, Rfv, respectively) were used to assess model accuracy. All six MLR models showed a good performance (Rc ≥ 0.90, Rcv ≥ 0.87, Rfv ≥ 0.81), with that for IMF content at the 2nd/3rd last rib showing the best predictive ability (Rc = 0.96, Rcv = 0.95, Rfv = 0.83). The Gabor filter-MLR models were applied to each pixel in an image to successfully derive distribution maps and visualize IMF content distribution in the loins. Results demonstrated that the conventional determination of IMF content of pork at different ribs along a single Longissimus dorsi could be replaced by pattern analysis technique-processed hyperspectral images of rib ends.

Near-infrared reflection spectroscopy and partial least squares regression to predict α-farnesene and conjugated trienol content in apples during storage

The synthesis of α-farnesene and its degradation to the corresponding oxidation products such as conjugated trienols (CTols) in apple skins is strongly correlated with the incidence of superficial scald, a physiological disorder occurring in apple during and after storage. This study aimed to evaluate the potential of near-infrared spectroscopy (NIRS) for the prediction of α-farnesene and CTols (CT258 and CT281) content, developing a rapid and non-invasive method to support the postharvest decision systems. Applying partial least squares (PLS) regression, positive correlations were found for α-farnesene and CTols obtaining correlation coefficients of calibration (rcal) above 0.90. The calibration model for one apple variety was validated with apples from a second season resulting in correlation coefficients of validation (rval) of 0.51, 0.71 and 0.76 for α-farnesene, CT258 and CT281, respectively. A global calibration model including two cultivars and two growing seasons led to standard errors of prediction (SEP) of 139, 60 and 59μmolm−2 for α-farnesene, CT258 and CT281, respectively. These results demonstrate the potential of NIRS for rapid and non-destructive prediction of the scald-disorder-related compounds α-farnesene and CTols in apple fruit.

High-throughput prediction of tablet weight and trimethoprim content of compound sulfamethoxazole tablets for controlling the uniformity of dosage units by NIR

This paper establishes a novel method to simultaneously predict the tablet weight (TW) and trimethoprim (TMP) content of compound sulfamethoxazole tablets (SMZCO) by near infrared (NIR) spectroscopy with partial least squares (PLS) regression for controlling the uniformity of dosage units (UODU). The NIR spectra for 257 samples were measured using the optimized parameter values and pretreated using the optimized chemometric techniques. After the outliers were ignored, two PLS models for predicting TW and TMP content were respectively established by using the selected spectral sub-ranges and the reference values. The TW model reaches the correlation coefficient of calibration (Rc) 0.9543 and the TMP content model has the Rc 0.9205. The experimental results indicate that this strategy expands the NIR application in controlling UODU, especially in the high-throughput and rapid analysis of TWs and contents of the compound pharmaceutical tablets, and may be an important complement to the common NIR on-line analytical method for pharmaceutical tablets.

Global and local QSPR models to predict supercooled vapour pressure for organic compounds

In this study, a quantitative structure–property relationship (QSPR) approach was used for estimation of logarithmic values of supercooled liquid vapour pressure (log PL) of a large set of structurally diverse organic compounds. This set includes 12 local sets of aromatic and aliphatic hydrocarbons, polychlorinated biphenyls, ethers, polychlorinated and brominated diphenylethers, polychlorinated naphthalenes and alcohols. Some simple models based on the linear relationship between log PL and VolSurf descriptors were developed as global models, and a general equation as a simple way to calculate the supercooled liquid vapour pressure of organic chemicals was provided. A descriptor representing the hydrophilic regions (WO1) of organic chemicals showed the highest correlation with log PL and resulted in a one-parameter global model characterized by satisfactory statistical performance; calibration (r2c) and prediction (r2p) correlation coefficient of 0.84 and 0.85, respectively. Moreover, local QSPR models were also developed for each subset of organic compounds and, as expected, the statistical results obtained from these models were better than the global one. From the descriptors involved in the models, it is concluded that the hydrophilic and hydrophobic regions at different energy levels and polarizability usually determine the variation of supercooled liquid vapour pressure of organic compounds.

Development of Calibration Models for Rapid Determination of Chemical Composition of Pacific Oyster (Crassostrea gigas) by Near Infrared Reflectance Spectroscopy

ABSTRACT The feasibility for rapid determination of the chemical composition of Pacific oyster (Crassostrea gigas) by near IR reflectance spectroscopy (NIRS) was assessed. NIRS calibration models were developed using 118 sets of C. gigas samples (two to three sets per batch, total of 54 batches, 5–10 individuals per set) collected over 10 mo from seven coastal locations in northeast China. Glycogen, protein, fat, taurine, zinc, selenium, and ash contents of gonad-visceral mass and adductor muscle were analyzed by laboratory-based methods. Freeze-dried samples were scanned by NIRS at 10,000–4,000 cm-1 and their chemical composition was predicted by partial least squares regression. Results showed that the correlation coefficient of calibration (RC) was greater than 0.90 for most chemical components tested (except taurine, 0.8266). Similar differences were found in the correlation coefficient and other parameter indices in both cross and external validations for various chemical components. This study indic...

Determination of Soluble Solid Content in Strawberry Using Hyperspectral Imaging Combined with Feature Extraction Methods

Hyperspectral imaging combined with feature extraction methods were applied to determine soluble sugar content (SSC) in mature and scatheless strawberry. Hyperspectral images of 154 strawberries covering the spectral range of 874-1,734 nm were captured and the spectral data were extracted from the hyperspectral images, and the spectra of 941~1,612 nm were preprocessed by moving average (MA). Nineteen samples were defined as outliers by the residual method, and the remaining 135 samples were divided into the calibration set (n = 90) and the prediction set (n = 45). Successive projections algorithm (SPA), genetic algorithm partial least squares (GAPLS) combined with SPA, weighted regression coefficient (Bw) and competitive adaptive reweighted sampling (CARS) were applied to select 14, 17, 24 and 25 effective wavelengths, respectively. Principal component analysis (PCA) and wavelet transform (WT) were applied to extract feature information with 20 and 58 features, respectively. PLS models were built based on the full spectra, the effective wavelengths and the features, respectively. All PLS models obtained good results. PLS models using full-spectra and features extracted by WT obtained the best results with correlation coefficient of calibration (r(c)) and correlation coefficient of prediction (r(p)) over 0.9. The overall results indicated that hyperspectral imaging combined with feature extraction methods could be used for detection of SSC in strawberry.