A strawberry quality detection method based on multiple light scattering and light absorption

Abstract

The present application relates to the field of strawberry quality detection, and discloses a kind of strawberry pulp quality detection method based on multiple light scattering and light absorption, first adopt optical integrating sphere system in combination with reverse doubling integral method, systematic determination different multi-species and multi-harvest storage period under the key quality index of strawberry pulp, optical absorption and scattering characteristic signal parameter, by optical characteristic pretreatment, utilize partial least square regression algorithm to construct the optical correlation parameter system of strawberry during storage physicochemical quality, realize by the prediction of the physicochemical quality of strawberry pulp optical characteristics after storage.This application first determines the light absorption and light scattering characteristic response signal of the physicochemical quality of strawberry during storage, and innovatively develops a rapid evaluation method for the physicochemical quality of strawberry pulp.The present application can realize the rapid prediction of the physicochemical quality of strawberry pulp during storage, and provide a time-saving and cost-controllable quality management method for the sorting, storage and transportation of planting enterprises after strawberry picking.

Description

Technical Field

[0001] This invention belongs to the field of spectral analysis technology, and relates to the application of visible-near-infrared spectroscopy to predict the quality of strawberries during storage. Background Technology

[0002] Strawberries, with their unique flavor and excellent nutritional value, are among the world's most popular fruits. However, the lack of a protective peel makes their skin extremely delicate and susceptible to mechanical damage. Furthermore, strawberries continue to experience high respiration and transpiration after harvest, leading to rapid metabolic depletion. During storage, strawberries undergo multiple quality changes, including skin discoloration, softening, nutrient loss, and susceptibility to microbial decay. These rapid and diverse changes significantly limit the market circulation and consumer acceptance of strawberries. Traditional physicochemical testing methods (such as texture analysis and chemical titration) can accurately reflect quality changes, but they are mostly destructive and time-consuming, making it difficult to achieve real-time, non-destructive monitoring of strawberry quality during storage and failing to meet the precision management needs of modern fruit and vegetable storage and transportation. Therefore, developing rapid, non-destructive monitoring methods is crucial for extending shelf life, optimizing storage conditions, and ensuring strawberry quality throughout the entire supply chain. Visible light (Vis) and near-infrared (NIR) spectroscopy, as efficient, rapid, and non-destructive detection methods, have been widely used in fruit and vegetable quality assessment. Currently, Vis-NIR spectroscopy can be used to rapidly predict the chemical quality of strawberries (such as soluble solids and titratable acidity). However, the accuracy of spectral-based detection of strawberry texture and microstructure remains limited, especially during storage. During storage, strawberry tissue structure undergoes complex and dynamic changes, and light experiences multiple scattering and absorption processes within the fruit, leading to insufficient predictive stability and poor model generalization ability. This reflects the limitations of Vis / NIR spectroscopy in assessing strawberry quality: composite optical information often masks the scattering and absorption characteristics of light within strawberry tissue.

[0003] Light propagates through strawberry pulp tissue primarily through two processes: absorption and scattering. μa mainly reflects changes in the chemical components of the fruit, such as pigments and water, while μs′ characterizes the cellular structure and tissue density. Therefore, systematically studying the changes in light absorption and scattering characteristics of strawberries during storage, along with their physicochemical quality, and revealing the intrinsic correlation mechanism between optical parameters and fruit quality, is of significant scientific and practical value for establishing a rapid prediction model for strawberry storage quality based on optical properties and achieving precise dynamic monitoring of strawberry quality. Summary of the Invention

[0004] The purpose of this invention is to provide a strawberry quality detection method based on multiple light scattering and light absorption, so as to predict the physicochemical quality of strawberry after storage based on the optical properties of strawberry tissue.

[0005] To achieve the above objectives, the present invention provides the following method:

[0006] This invention provides a strawberry quality detection method based on multiple light scattering and light absorption:

[0007] S1: Select two types of 80% ripe and intact strawberries with similar appearance and size, store them under low temperature and constant humidity, and take samples at different storage points to determine their physicochemical quality and optical properties.

[0008] S2: Strawberry slices were cut and their transmittance and reflectance in the visible-near infrared band were measured using a single integrating sphere optical measurement system. The optical characteristic parameters were obtained using the IAD code.

[0009] S3: The color of strawberries was measured using a colorimeter, the hardness was measured using a texture analyzer, the soluble solids content was determined using a digital refractometer, the total acidity was determined using an acidity meter, and the pectin content was determined after cell wall material was extracted using the alcohol-soluble method.

[0010] S4: After preprocessing the optical property parameters by standard normal transformation and smoothing, partial least squares cross-validation is used to fit the spectral and physicochemical quality parameters to the model. After training, the cross-validation model is obtained, which can be used to predict and evaluate the physicochemical quality of strawberries.

[0011] In a preferred embodiment of the present invention, in S1, 105 freshly picked "Xiangye" and "Hongyan" strawberry samples were selected and stored in an environment of 5±2℃ and 80% relative humidity, for a total of 210 strawberry samples. At 0, 2, 4, 6, and 8 days, 15 strawberry samples of each of the two varieties were randomly selected each day to determine their physicochemical quality, including L... * a * b * Hardness, soluble solids (SSC), titratable acid (TA) content, water-soluble pectin (WSP), chelated pectin (CSP), and alkali-soluble pectin (NSP).

[0012] In a preferred embodiment of the present invention, in step S2, a fresh strawberry sample is cut into 2 mm thick slices and placed between two quartz glass slides (1 mm thick, refractive index = 1.53). Using a single integrating sphere optical measurement system, the absorption coefficient (μ) of the strawberry sample in the visible-near infrared (Vis-NIR, 400-1050 nm) band is measured. a ) and reduced scattering coefficient (μ s At 0, 2, 4, 6, and 8 days, six strawberry samples from each of two varieties were measured daily, resulting in a total of 120 spectral data points (5 storage times × 2 varieties × 6 replicates × 2 optical characteristics (μ)). a / μ s ′)).

[0013] In a preferred embodiment of the present invention, in step S3, a colorimeter is used to randomly select two symmetrical points at the equatorial center of each strawberry for measurement, and L is taken as the value. * a * b * The average value was used as the representative value for each strawberry sample. Using a texture analyzer, measurements were taken at two symmetrical points along the equatorial center line of each strawberry, and the average hardness was calculated as the representative measurement value for each strawberry sample. After removing the strawberry sepals, the strawberry fruit was cut into pieces, and a portion of the strawberry pieces were squeezed through double-layered gauze to obtain clear strawberry juice. The strawberry juice was analyzed using a digital refractometer to determine the SSC (saturated chromatogram), and the measurement was repeated twice, with the average value used as the representative value for each strawberry sample. Subsequently, the strawberry juice was diluted 50 times with pure water, and the TA (total acidity) was determined using a pH meter. This measurement was repeated twice, with the average value used as the representative value for each strawberry sample. The remaining strawberry pieces were collected for determining pectin content. Cell wall material was extracted from the remaining strawberry pieces after physicochemical quality measurements using an alcohol-soluble method. WSP (whole cell wall protein), CSP (cell wall protein), and NSP (non-stress protein) were extracted sequentially. Measurements were repeated 6 times at each storage stage; a total of 60 sets of physicochemical qualities corresponding to the spectra were obtained (5 storage times × 2 varieties × 6 replicates).

[0014] In a preferred embodiment of the present invention, in step S4, the 120 spectral data points undergo standard normal variable (SNV) and smoothing (Savitzky-Golay derivative) preprocessing. The 120 spectra (μ a / μ s All samples were used as the training set. A 10-fold cross-validation method was used to divide the samples into 10 groups, and each group was used as the validation set for model fitting, resulting in 10 candidate models. The root mean square error (RMSECV) of cross-validation for each candidate model was calculated, and the solution with the minimum value was selected to determine the model weights. Using Matlab software and the Partial Least Squares Regression (PLS) toolbox, a quality prediction model for strawberries during storage based on optical properties was established. The model was then used to correct the root mean square error (RMSECV), RMSECV, and the prediction accuracy of the dataset (R²). c 2 ) and cross-validation dataset prediction accuracy (R cv 2 The predictive performance of the PLS model is evaluated using the coefficient of determination and the number of latent variables (LVs).

[0015] The advantages of this invention are:

[0016] (1) The optical response mechanism of the physicochemical quality of strawberry fruit during storage was clarified, providing an important theoretical basis for the optical detection of strawberry fruit. (2) A method for rapidly predicting the physicochemical quality of strawberries was developed. Compared with traditional manual detection, this invention saves time, reduces labor, and significantly reduces the cost of testing. (3) By rapidly characterizing the optical properties of strawberry pulp, the physicochemical quality of strawberries during storage can be predicted, providing controllable quality management for strawberry growing enterprises in the sorting, storage, and transportation stages after strawberry harvesting. Attached Figure Description

[0017] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the accompanying drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. In all the drawings, similar elements or parts are generally identified by similar reference numerals. In the drawings, the elements or parts are not necessarily drawn to scale. Attached image description:

[0019] Figure 1 Abstract with attached figures

[0020] Figure 2 Changes in appearance of "Xiangye" and "Hongyan" strawberries during storage

[0021] Figure 3 Correlation of physicochemical qualities of strawberries during storage (A. Xiangye; B. Hongyan)

[0022] Figure 4 Optical properties of strawberries during storage (A. μ-values ​​of Xiangye strawberry) a B. μ of the fragrant wild strawberry s ′;C. μ of Red Beauty Strawberry a D. μ of the red-faced strawberry s ′)

[0023] Figure 5 Correlation between physicochemical quality and optical properties of strawberries during storage (A. μ-values ​​of strawberry). a With L * a * b * A. Correlation between SSC and TA; B. Correlation between *Strawberry fragrantii* and firmness, WSP, CSP, and NSP (μ) s ′;C. μ of Red Beauty Strawberry a With L * a * b * The correlation between SSC and TA; D. μ of Red Beauty strawberry s ′) Detailed Implementation

[0024] 1. Experimental Materials and Methods:

[0025] Strawberry Harvesting and Storage: "Xiangye" and "Hongyan" strawberries were harvested from two separate plantations and transported to the laboratory on the day of harvest. After harvesting, the fruits were stored at 5 ± 2 ℃ and 80% relative humidity. A total of 105 fruits of each variety were collected, with 30 samples used for optical characterization and the remaining 75 samples used for physicochemical quality determination. Optical characteristics and physicochemical indicators were measured at five storage time points (0, 2, 4, 6, and 8 days). Changes in appearance during strawberry storage are shown below. Figure 1 As shown.

[0026] Optical property acquisition of strawberries: Measurements were performed using a single integrating sphere system. The apparatus included a halogen lamp (ASBN-W100-L, Spectral Products, USA), an integrating sphere (4P-GPS-033-SL, Labsphere, USA), and a visible-near-infrared spectrometer with a wavelength range of 400-1050 nm. A visible-near-infrared spectrometer (PG2000-PRO, Shanghai Aidi Optoelectronics Co., Ltd., China) with wavelength range of 400-1050 nm was also used, along with a computer equipped with spectral analysis software. Measurements were performed in a dark environment. Fresh strawberry samples were cut into 2 mm thick slices and placed between two quartz glass slides (1 mm thick, refractive index = 1.53) for optical testing. Reflectance (Rs, Rd, Rt) and transmittance (Ts, Td, Tt) signals were obtained by measuring in both reflectance and transmittance modes of the spectrometer. μ was derived from the measured reflectance (R) and transmittance (T) data using the inverse doubling algorithm (IAD). a and μ s ′.

[0027] The reference formula is as follows:

[0028]

[0029]

[0030] Physicochemical quality determination: The color of the strawberry equatorial region was measured using a digital colorimeter (CR-400, Konica Minolta Sensors Co., Ltd., Tokyo, Japan). Two symmetrical points were randomly selected at the center of the equator of each strawberry for measurement, and the average value was taken as the representative value of the sample. After removing the strawberry sepals, the fruit was squeezed through double-layered gauze to obtain clear juice. The juice was analyzed using a digital refractometer (PAL-1, Atago, Tokyo, Japan) to determine the soluble solids content (SSC). Subsequently, the juice was diluted 50 times with purified water, and the titratable acid (TA) was determined using a PAL-Easy Acid F5 pH meter (PAL-Easy Acid F5, Atago, Tokyo, Japan). The hardness was measured at two symmetrical points on the center line of the equator of each strawberry using a texture analyzer (TMS-Pro, FTC, USA), and the average value was calculated as the representative measurement value of each sample. The specific parameter settings were as follows: cylindrical probe diameter 5 mm, maximum load sensor value 50 N, initial force 0.4 N, and puncture depth 5 mm. This study used the method described by De Roeck et al. (2008) to determine pectin, and adjusted the specific parameters. First, alcohol-insoluble residue (AIR) was prepared, and then water-soluble pectin (WSP), chelating agent-soluble pectin (CSP), and sodium carbonate-soluble pectin (NSP) were extracted sequentially from the obtained AIR.

[0031] Model Establishment: A partial least squares regression (PLSR) model for predicting the physicochemical quality of strawberries was constructed using the PLS toolbox (Eigenvector Research Inc., Winnage, USA) in MATLAB® (R2023a, MathWorks, USA). 10-fold cross-validation was employed. Preprocessing of the spectral data used standard normal variables (SNV) and smoothing (Savitzky-Golay method, interval size = 15, polynomial order = 2) to achieve optimal predictive performance. The model was further refined by correcting for root mean square error (RMSEC), RMSECV, and the predicted accuracy of the dataset (R²). c 2 ) and cross-validation dataset prediction accuracy (R cv 2 The predictive performance of the PLS model is evaluated using the coefficient of determination and the number of latent variables (LVs).

[0032] 2. Changes in the physicochemical quality of strawberries during storage:

[0033] Table 1 shows the changes in the physicochemical quality of strawberries during storage. The L values ​​of the two strawberry varieties during storage are also shown. * The values ​​all showed a downward trend, with the L value of "Xiangye" strawberry showing a downward trend. * The value decreased significantly (P < 0.05). * The values ​​all showed an upward trend, with the "Xiangye" strawberry's a* The value fluctuated upwards, while the "red beauty" strawberry's a * The value increased significantly (P < 0.05). The b values ​​of the two strawberry varieties increased significantly. * All values ​​showed a fluctuating downward trend, with the decrease being more significant for "Xiangye" strawberries (P < 0.05). The firmness of both "Xiangye" and "Hongyan" strawberries decreased significantly (P < 0.05). The firmness of "Xiangye" strawberries decreased from 3.60 to 2.53 throughout the storage period, while the firmness of "Hongyan" strawberries decreased from 2.82 to 1.77. During storage, both the SSC and TA of "Xiangye" strawberries showed a downward trend. However, the decrease in TA was significant (from 0.40 to 0.31; P < 0.05), while the change in SSC was not significant (from 10.75 to 9.95). The SSC and TA of "Hongyan" strawberries also showed a downward trend, with TA decreasing significantly (P < 0.05) from 0.47 to 0.35, and SSC decreasing from 11.32 to 10.92. The WSP of 'Xiangye' strawberries significantly increased from day 0 to 4 (P < 0.05), rising from 81.86 to 105.25 g GalA·kg⁻¹, before slightly decreasing to 99.44 g GalA·kg⁻¹ by day 8. In contrast, the WSP of 'Hongyan' strawberries showed a continuous downward trend, decreasing from 319.52 to 187.14 g GalA·kg⁻¹. The CSP content of 'Xiangye' strawberries significantly decreased from day 0 to 4 (P < 0.05), falling from 74.28 to 23.55 g GalA·kg⁻¹, before rebounding to 45.29 g GalA·kg⁻¹ by day 8. However, the CSP content of 'Hongyan' strawberries showed a continuous downward trend, decreasing from 246.29 g GalA·kg⁻¹ to 154.63 g GalA·kg⁻¹. The NSP content of both types of strawberries showed a decreasing trend during storage. Specifically, the NSP content of "Xiangye" decreased from 37.27 g GalA·kg⁻¹ to 22.95 g GalA·kg⁻¹, and that of "Hongyan" decreased from 134.57 g GalA·kg⁻¹ to 82.44 g GalA·kg⁻¹.

[0034] Table 1. Changes in the physicochemical quality of "Xiangye" and "Hongyan" strawberries during storage.

[0035]

[0036] Note: Different letters indicate significance between different storage times, P < 0.05

[0037] 3. Changes in optical properties and their correlation with physicochemical properties:

[0038] Figure 4 and Figure 5The optical properties of 'Xiangye' and 'Hongyan' strawberries in the 450–1050 nm wavelength range during storage and their correlation coefficients with their physicochemical quality are presented. For μ... a The |r| value of the "Xiangye" strawberry in the 960-1050 nm band is higher than that in the 450-600 nm band. * Value and μ a Overall, there is a positive correlation, with an average correlation coefficient r of 0.85 in the 960–1050 nm band. Conversely, L * SSC and TA with μ a They showed a negative correlation, with average correlation coefficients r of -0.94, -0.87, and -0.91 in the same band, respectively. * Value and μ a A weak correlation was observed. For the "Red Beauty" strawberry variety, except for L... * and b * Furthermore, the average |r| values ​​in the 450–600 nm and 960–1050 nm bands are relatively close. * Value and μ a A positive correlation was observed, with average r values ​​of 0.86 in both the 450–600 nm and 960–1050 nm wavelength bands. Soluble solids content (SSC) and total acidity (TA) were correlated with μ... a They are negatively correlated, with average r values ​​of -0.95 and -0.93 for the two bands, respectively. * The |r| value is higher in the 450–600 nm band (average r = −0.86) than in the 960–1050 nm band, while b * With μ a A weak correlation was observed. The μ value of strawberries... a The value is mainly determined by moisture content, in addition, Figure 4 The absorption peak near 980 nm is attributed to the third harmonic of the O–H stretching vibration. Therefore, the |r| values ​​in the 960–1050 nm band are generally higher than those in the 450–600 nm band. Notably, in the 600–960 nm band, the |r| value of the "Xiangye" strawberry changes drastically, while the |r| value of the "Hongyan" strawberry remains relatively stable, indicating that this band is more sensitive to variety-specific differences in strawberries.

[0039] And for μ s ′, Figure 5 B and D show the fruit firmness, AIR, WSP, CSP, NSP, and μ values ​​of two strawberry varieties in the 450–1050 nm wavelength range. sThe correlation coefficient r between the two values ​​is shown. Within the 600–1000 nm wavelength range, the correlation is relatively stable with a high |r| value. The average r values ​​for firmness, AIR, WSP, CSP, and NSP of the 'Xiangye' strawberry are 0.99, −0.63, −0.77, 0.73, and 0.81, respectively. The corresponding average r values ​​for 'Hongyan' are 0.89, 0.90, 0.93, 0.79, and −0.72. The high correlation between fruit firmness and optical properties stems from tissue microstructure and cell wall polysaccharide characteristics, especially cell wall polysaccharide content, cell number, and porosity.

[0040] 4. Prediction Model Establishment

[0041] To further explore the relationship between spectral parameters and the physicochemical quality of strawberries during storage, this study is based on the μ-strain of strawberries. a and μ s A partial least squares regression (PLSR) model was established based on the spectral parameters. a The spectral model showed good accuracy in predicting physicochemical properties, including L. * Value (R) cv ² = 0.93, RMSECV = 0.89), a * Value (R) cv ² = 0.81, RMSECV = 1.39), hardness (R cv ² =0.89, RMSECV = 0.22), WSP (R cv ² = 0.85, RMSECV = 35.52), CSP (R cv ² = 0.84, RMSECV = 31.92) and NSP(R cv ² = 0.88, RMSECV = 15.12) both showed good predictive performance. In contrast, b * (R) cv ² =0.75, RMSECV = 2.21), SSC (R cv ² = 0.61, RMSECV = 0.83) and TA(R cv ² The predictive performance of μ (= 0.63, RMSECV = 0.05) is relatively weak. s The predictive power of ′ is slightly lower than that of μ. a Overall, the model performed better in predicting color and pectin content, but had lower accuracy in predicting quality parameters such as SSC and TA.

[0042] Table 2 μ based on 450–1050 nm a and μ s Establish a PLSR physicochemical quality prediction model

[0043]

[0044] LVs: Latent variables; R c 2 : Corrects the prediction accuracy of the dataset; RMSEC: Corrects the root mean square error; R cv 2 : Prediction accuracy of the cross-validation dataset; RMSECV: Root mean square error of cross-validation.

Claims

1. A method for detecting strawberry quality based on multiple light scattering and light absorption, characterized in that, The method includes: S1: Select two types of 80% ripe and intact strawberries with similar appearance and size, store them under low temperature and constant humidity, and take samples at different storage points to determine their physicochemical quality and optical properties. 2.S2: Strawberry slices were prepared and their transmittance and reflectance in the visible-near-infrared band were measured using a single integrating sphere optical measurement system. The optical characteristic parameters were obtained using the IAD code. S3: The color of strawberries was measured using a colorimeter, the hardness was measured using a texture analyzer, the soluble solids content was determined using a digital refractometer, the total acidity was determined using an acidity meter, and the pectin content was determined after cell wall material was extracted using the alcohol-soluble method. 3.S4: After preprocessing the optical characteristic parameters by standard normal transformation and smoothing, partial least squares cross-validation is used to fit the spectral and physicochemical quality parameters to the model. After training, the cross-validation model is obtained, which can be used to predict and evaluate the physicochemical quality of strawberries.

4. The strawberry quality detection method based on multiple light scattering and light absorption according to claim 1, characterized in that... As described in S1: 105 freshly picked "Xiangye" and "Hongyan" strawberry samples were selected and stored in an environment of 5±2℃ and 80% relative humidity, for a total of 210 strawberry samples. At 0, 2, 4, 6, and 8 days, 15 strawberry samples of each of the two varieties were randomly selected daily to determine their physicochemical quality, including L... * a * b * Hardness, soluble solids (SSC), titratable acid (TA) content, water-soluble pectin (WSP), chelated pectin (CSP), and alkali-soluble pectin (NSP).

5. The strawberry quality detection method based on multiple light scattering and light absorption according to claim 1, The feature is that, as described in S2, a fresh strawberry sample is cut into 2 mm thick slices and placed between two quartz glass slides (1 mm thick, refractive index = 1.53). The absorption coefficient (μ) of strawberry samples in the visible-near infrared (Vis-NIR, 400-1050 nm) band was determined using a single integrating sphere optical measurement system. a ) and reduced scattering coefficient (μ s At 0, 2, 4, 6, and 8 days, six strawberry samples from each of two varieties were measured daily, resulting in a total of 120 spectral data points (5 storage times × 2 varieties × 6 replicates × 2 optical characteristics (μ)). a / μ s ′)).

6. The strawberry quality detection method based on multiple light scattering and light absorption according to claim 1, characterized in that... As described in S3: Using a colorimeter, two symmetrical points are randomly selected at the equatorial center of each strawberry for measurement, and L is taken as the measurement point. * a * b * The average value was used as the representative value for each strawberry sample. Using a texture analyzer, measurements were taken at two symmetrical points along the equatorial center line of each strawberry, and the average firmness was calculated as the representative measurement value for each strawberry sample. After removing the strawberry sepals, the strawberry fruit was cut into pieces, and a portion of the strawberry pieces were squeezed through double-layered gauze to obtain clear strawberry juice. The strawberry juice was analyzed using a digital refractometer to determine the SSC (saturated chromatogram), and the measurement was repeated twice, with the average value taken as the representative value for each strawberry sample. Subsequently, the strawberry juice was diluted 50 times with pure water, and the TA (total acidity) was determined using a pH meter. This measurement was repeated twice, with the average value taken as the representative value for each strawberry sample. The remaining strawberry pieces were collected for determining pectin content. Cell wall material was extracted from the remaining strawberry pieces after physicochemical quality measurements using an alcohol-soluble method. WSP (whole cell wall protein), CSP (cell wall protein), and NSP (non-stress protein) were extracted sequentially. Measurements were repeated 6 times at each storage stage; a total of 60 sets of physicochemical qualities corresponding to the spectra were obtained (5 storage times × 2 varieties × 6 replicates).

7. The strawberry quality detection method based on multiple light scattering and light absorption according to claim 1, characterized in that... As described in S4: Standard normal variable (SNV) and smoothing (Savitzky-Golayderivative) preprocessing was performed on 120 spectral data points. 120 spectra (μ a / μ s All samples were used as the training set. A 10-fold cross-validation method was used to divide the samples into 10 groups, and each group was used as the validation set for model fitting, resulting in 10 candidate models. The root mean square error (RMSECV) of cross-validation for each candidate model was calculated, and the solution with the minimum value was selected to determine the model weights. Using Matlab software and the Partial Least Squares Regression (PLS) toolbox, a quality prediction model for strawberries during storage based on optical properties was established. The model was then used to correct the root mean square error (RMSECV), RMSECV, and the prediction accuracy of the dataset (R²). c 2 ) and cross-validation dataset prediction accuracy (R cv 2 The predictive performance of the PLS model is evaluated using the coefficient of determination and the number of latent variables (LVs).

Images