A non-destructive determination method for peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction

Abstract

The present application discloses a non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction, belonging to the field of spectral technology detection. First, olive oil samples with different oxidation degrees are obtained and the standard values of peroxide value and acid value are tested; absorption spectra and laser-induced fluorescence spectra are collected; candidate characteristic fluorescence wavelengths are determined through spectral analysis; the influence of fluorescence inner filter effect on the detection results under different candidate characteristic fluorescence wavelengths is verified to obtain the optimal characteristic fluorescence wavelength; a quantitative prediction model for the peroxide value and acid value of olive oil is established with the fluorescence intensity and absorption coefficient at the optimal characteristic fluorescence wavelength, the absorption coefficient at the excitation wavelength, and the standard values of peroxide value and acid value. The modeling of the present invention only requires absorption signals at two wavelengths and a fluorescence signal at a single wavelength, which can effectively eliminate the influence of fluorescence inner filter effect on the detection results and achieve accurate, rapid and non-destructive prediction of the peroxide value and acid value of different brands of olive oil.

Description

Technical Field

[0001] This application relates to the field of spectroscopic technology detection, in particular to a non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction. Background Technique

[0002] Lipid oxidation is one of the main aspects leading to the deterioration during the storage and processing of edible oils. Oxidation seriously affects its quality, such as changes in color, flavor, aroma and nutritional value. Virgin olive oil is a vegetable oil obtained from olive fruits through physical cold pressing. Due to its advantages of rich antioxidants, high oleic acid ratio, high monounsaturated fat, low polyunsaturated fat, low acidity, good sensory properties and positive effects on the cardiovascular system, it is popular among consumers and has a relatively high price compared to other vegetable oils. Although olive oil has high oxidation stability, it is also easily affected by the oxidation process. Environmental storage conditions, such as heat, light and oxygen exposure, can promote the oxidation process. Autoxidation also occurs due to the changes in the inherent compounds in virgin olive oil. The oxidation of virgin olive oil leads to the formation of primary oxidation products, such as hydroperoxides, which further decompose into secondary oxidation products, such as ketones, aldehydes and alcohols, resulting in rancidity (off-flavor).

[0003] The peroxide value (PV) and acid value (AV) are the main indicators for evaluating the oxidative rancidity state of virgin olive oil. Currently, the national standard GB5009.229-2016 stipulates that the determination methods for the acid value of vegetable oils include: cold solvent indicator titration method (the first method), cold solvent automatic potentiometric titration method (the second method), and hot ethanol indicator titration method (the third method). The national standard GB5009.227-2016 stipulates that the determination method for the peroxide value of vegetable oils is: titration method (the first method), potentiometric titration method (the second method). The national standard method has high specificity and accurate detection, but the sample pretreatment and titration process are cumbersome, with a long cycle, and a large number of reagents are required, posing potential risks to the health of operators and the environment. In recent years, there have been relatively many reports on the rapid non-destructive detection of vegetable oil oxidation, mainly including: electronic nose technology, near-infrared spectroscopy technology, Fourier transform infrared spectroscopy technology, Raman spectroscopy technology and fluorescence spectroscopy technology. However, the previous several technologies are not very sensitive to the detection of low-concentration components and require complex spectral pretreatment, establishing complex models through chemometrics or machine learning. Laser-induced fluorescence spectroscopy technology has higher sensitivity and selectivity for organic and inorganic compounds, generally does not require the use of consumable reagents and sample pretreatment, has relatively obvious characteristic fluorescence peaks, and can achieve signal enhancement for target fluorescent substances through laser induction.

[0004] At present, the detection applications of fluorescence spectroscopy technology in the safety and quality of vegetable oils mainly include: (1) quality grading; (2) adulteration monitoring; (3) origin traceability; (4) quantitative prediction of fluorescent components; (5) monitoring of thermal oxidation and photooxidation; (6) assessment of quality changes during storage. However, when using fluorescence spectroscopy technology to detect the related quality and safety of food or agricultural products, the detection effect is often affected by the complex matrix of the sample. This effect is mainly manifested as the fluorescence inner filter effect, that is, the non-linear relationship between the fluorescence intensity and the fluorophore concentration. This is because there are a large number of chromophores in the sample matrix that absorb light. The absorption of the excitation light and the fluorescence emission light by these chromophores results in the attenuation of the fluorescence intensity and the shift of the position of the fluorescence characteristic peak. The fluorescence inner filter effect includes: primary inner filter effect and secondary inner filter effect. The primary refers to the absorption of the excitation light, and the secondary, also known as reabsorption, refers to the absorption of the fluorescence. Therefore, developing a non-destructive determination method for the peroxide value and acid value of virgin olive oil that can eliminate the fluorescence inner filter effect is of great significance for ensuring the quality of virgin olive oil. Summary of the Invention

[0005] In order to eliminate the influence of the fluorescence inner filter effect, the present invention proposes a non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction, and establishes a highly robust calculation model for the peroxide value and acid value.

[0006] For the above purpose, the technical solution adopted by the present invention is:

[0007] A non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction, comprising the following steps:

[0008] S1. Prepare a virgin olive oil sample;

[0009] S2. Determine the peroxide value PV and acid value AV of the virgin olive oil sample, and use the determination results as the actual values;

[0010] S3. At a specific excitation wavelength, collect the absorption spectrum and fluorescence spectrum of the sample within a certain fluorescence wavelength range;

[0011] S4. Spectral analysis: According to the fluorescence spectrum waveforms of different brand olive oil samples, determine the candidate characteristic fluorescence wavelengths from the positions of the fluorescence peaks;

[0012] S5. Verify the influence of the fluorescence inner filter effect on the detection results at different candidate characteristic fluorescence wavelengths, and determine the optimal characteristic fluorescence wavelength;

[0013] S6. According to the fluorescence intensity F at the optimal characteristic fluorescence wavelength em , absorption coefficient μ a,em and the absorption coefficient μ at the excitation wavelength a,ex, using non - linear least - squares fitting, a quantitative prediction model for the peroxide value PV and acid value AV corrected based on the fluorescence inner filter effect is obtained:

[0014]

[0015]

[0016] Among them, y1 is the peroxide value PV, a1, b1, c1, d1, e1 are the fitting parameters of the quantitative prediction model of the peroxide value PV, y2 is the acid value AV, and a2, b2, c2, d2, e2 are the fitting parameters of the quantitative prediction model of the acid value AV;

[0017] S6. Collect the fluorescence intensity F em , absorption coefficient μ a,em and the absorption coefficient μ a,ex at the excitation wavelength, and use the quantitative prediction model of the peroxide value PV and acid value AV corrected based on the fluorescence inner filter effect established in step S6 to calculate the peroxide value and acid value of the virgin olive oil to be measured.

[0018] Further, in step S1, virgin olive oils of different brands are placed in brown bottles and subjected to accelerated oxidation treatment in an oven at 110 °C, and samples are taken at different time nodes and cooled for use.

[0019] Further, in step S3, a certain amount of olive oil samples are taken, and a tungsten - halogen lamp and a laser with a wavelength of 375 nm are used as light sources respectively, the integration time is 2 s for both, and the smoothing times are 2, to obtain the fluorescence spectra and absorption spectra of the olive oil samples in the wavelength range of 200 - 1000 nm.

[0020] Further, the candidate characteristic fluorescence wavelengths include 445 nm, 475 nm, 525 nm, 680 nm, and 725 nm.

[0021] Further, step S5 includes the following specific contents:

[0022] First, divide the virgin olive oil sample set of different brands, use one brand of olive oil as the validation set, and the remaining brand olive oils as the calibration set;

[0023] Using the fluorescence intensity F em at the candidate characteristic fluorescence wavelengths as the input parameter, and the peroxide value PV and acid value AV as the output parameters respectively, use the calibration set to establish the first support vector machine regression model for the peroxide value and acid value of olive oil;

[0024] And, using the absorption coefficient μ a,ex, the absorption coefficient μ at the fluorescence wavelength of the candidate feature a,em and the fluorescence intensity F em Taking the absorption coefficient μ and the fluorescence intensity F at the fluorescence wavelength of the candidate feature as input parameters, and the peroxide value PV and the acid value AV as output parameters respectively, a second support vector machine regression model for the peroxide value and acid value of olive oil is established by using the calibration set;

[0025] Use the validation set to test the effects of the first support vector machine regression model and the second support vector machine regression model at different fluorescence wavelengths of the candidate features. Take the correlation coefficient and the root mean square error as evaluation indicators, and determine the optimal fluorescence wavelength according to the test results.

[0026] Furthermore, the optimal fluorescence wavelength is 475 nm.

[0027] Furthermore, the quantitative prediction models of the peroxide value PV and the acid value AV corrected based on the fluorescence inner filter effect are as follows:

[0028]

[0029]

[0030] where F em and μ a,em are the fluorescence intensity and the absorption coefficient at 475 nm respectively, and μ a,ex is the absorption coefficient at 375 nm.

[0031] Based on the above content, a non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction proposed by the present invention has the following beneficial effects compared with the existing detection technology:

[0032] (1) Fast and non-destructive. Since the present invention realizes the detection of the peroxide value and acid value based on the fluorescence spectrum and absorption coefficient spectrum of virgin olive oil, no sample pretreatment is required, and the time consumed is only the spectrum acquisition time, about 6 s;

[0033] (2) Few parameters are required for prediction. Only three optical parameters are required: the absorption coefficients at 375 nm and 475 nm and the fluorescence intensity at 475 nm.

[0034] (3) The established prediction models for the peroxide value PV and the acid value AV have high accuracy and good robustness, and are applicable to the calculation of PV and AV of different brands of olive oil. In order to verify the robustness of the proposed model, when grouping the samples, one brand of olive oil was selected as the validation set, and the other brands of olive oil were used as the calibration set. Not all brands of olive oil were included during the model calibration. The results show that the established model has high accuracy in predicting the PV and AV of olive oil of brands other than the calibration set. Description of the Drawings

[0035] The accompanying drawings here are incorporated into the specification and form a part of this specification, showing embodiments consistent with this application, and are used together with the specification to explain the principles of this application.

[0036] Figure 1 It is the change trend of peroxide value and acid value measured by the national standard method during the thermal oxidation process of olive oil with oxidation time.

[0037] Figure 2 It is the absorption coefficient spectrum and fluorescence intensity spectrum of 5 brands of olive oil at different oxidation times.

[0038] Figure 3 It is the change trend of fluorescence intensity of 5 brands of olive oil at 475 nm with oxidation time.

[0039] Figure 4 It is the change trend of absorption coefficient of 5 brands of olive oil at 375 nm with oxidation time.

[0040] Figure 5 It is the change trend of absorption coefficient of 5 brands of olive oil at 475 nm with oxidation time.

[0041] Figure 6 It is the distribution diagram of the calculated peroxide value of olive oil samples in the calibration set and validation set with the actual value.

[0042] Figure 7 It is the distribution diagram of the calculated acid value of olive oil samples in the calibration set and validation set with the actual value.

[0043] Figure 8 It is the schematic flow diagram of a non-destructive determination method for peroxide value and acid value of olive oil corrected based on fluorescence inner filter effect. Detailed Embodiments

[0044] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0045] Figure 8 It shows a non-destructive determination method for peroxide value and acid value of olive oil corrected based on fluorescence inner filter effect proposed by the present invention. The method includes the following steps:

[0046] S1. Sample preparation:

[0047] In order to obtain various oxidized olive oils, accelerated oxidation treatment was carried out on 5 brands of olive oil. First, each kind of olive oil was respectively filled in brown bottles. Then, the olive oil samples were placed in an oven at a temperature of 110 °C and heated continuously for 96 h. Samples were taken every 12 hours, 10 samples (each sample was 30 mL) were taken for each kind of olive oil each time, and samples were taken 9 times. Each kind of olive oil had 90 samples, and there were a total of 450 olive oil samples.

[0048] S2. Determination of sample PV and AV:

[0049] The peroxide value (PV) of the sample was determined according to the first method of GB 5009.227-2016. Weigh 2.00 - 3.00 g of the sample and place it in a 250 mL iodine flask. Add 30 mL of a chloroform - glacial acetic acid mixed solution and gently shake until the sample is completely dissolved. Then, add 1.00 mL of saturated potassium iodide solution, tightly cap the bottle, and gently shake for 30 s. Immediately place it in a dark environment for 3 min. Finally, take out the sample, add 100 mL of water, shake well, and immediately titrate with a sodium thiosulfate standard solution until it turns light yellow. Then add 1.00 mL of starch indicator and continue titrating with strong shaking until the blue color of the solution disappears as the end point. At the same time, perform a blank test. The volume V0 of the 0.01 mol / L sodium thiosulfate solution consumed in the blank test should not exceed 0.1 mL. The calculation formula is as follows:

[0050]

[0051] Among them, PV is the peroxide value, with the unit of meq / kg; V is the volume of the sodium thiosulfate standard solution consumed by the sample, with the unit of mL; V0 is the volume of the sodium thiosulfate standard solution consumed in the blank test, with the unit of mL; c is the concentration of the sodium thiosulfate standard solution, with the unit of mol / L; m is the mass of the sample taken, with the unit of g.

[0052] The acid value (AV) of the sample was determined according to the first method of GB 5009.229-2016. Weigh 3.00 - 5.00 g of the sample and place it in a 250 mL iodine flask. Add 100 mL of an ether - isopropanol mixed solution and 3 drops of phenolphthalein indicator, and titrate with a sodium hydroxide standard titration solution. When the solution turns light pink and does not fade within 15 s, it is the end point of the titration. The calculation formula is as follows:

[0053]

[0054] Among them, AV is the acid value, with the unit of mg / g; V is the volume of the standard titration solution consumed in the sample determination, with the unit of mL; V0 is the volume of the standard titration solution consumed in the corresponding blank determination, with the unit of mL; c is the molar concentration of the standard titration solution, with the unit of mol / L; m is the mass of the sample taken, with the unit of g.

[0055] The changing trends of PV and AV of the 5 brands of olive oil measured with the oxidation time are as Figure 1As shown in the figure, it can be found that the PV values of the five brands of olive oil all show an obvious upward trend with the heating time, but the increasing degrees and the final PV and AV values of the five brands of olive oil are different. The PV value range of the 450 prepared olive oil samples is 4.627 - 58.910 meq / kg, and the AV value range is 0.353 - 1.406 mg / g.

[0056] S3. Absorption spectrum and fluorescence spectrum acquisition:

[0057] The absorption coefficient and fluorescence intensity of the olive oil samples are obtained through an independently developed optical platform (patent application number CN202011238253.4). Take a certain amount of olive oil sample, use a tungsten halogen lamp and a laser (375 nm) as light sources in the double integrating sphere module and the laser-induced fluorescence module respectively, set the integration time to 2 s and the smoothing times to 2, and obtain the transmittance-reflectance spectrum and fluorescence intensity spectrum of the olive oil in the wavelength range of 200 - 1000 nm. Further, the absorption and reduced scattering coefficients are calculated from the transmittance-reflectance spectrum using the inverse adding-doubling (IAD) algorithm. Since this invention is for the calculation of PV and AV of different brands of olive oil, the difference in the reduced scattering coefficient is small, while the difference in the absorption coefficient is large, and the fluorescence inner filter effect has a greater impact on the calculation results. Therefore, only the absorption coefficient and fluorescence intensity need to be obtained.

[0058] S4. Spectral analysis:

[0059] The μ a spectrum and fluorescence intensity spectrum of different brands of olive oil after heat oxidation treatment are as Figure 2 shown. The main fluorescence peaks of the five different brands of olive oil are: 445, 475, 525, 680, 725 nm. Among them, 445 and 475 nm are related to the oxidation products of fatty acids and vitamin E, and the peak at 525 nm originates from vitamin E. The peak intensities at 445 and 475 nm are very low because virgin olive oil has a high content of monounsaturated fatty acids and phenolic antioxidants, which can provide greater antioxidant protection. The fluorescence intensities at these characteristic fluorescence wavelengths show an obvious downward or upward trend with the oxidation time. Therefore, for a single brand, the fluorescence intensities at these characteristic wavelengths can be used for PV and AV calculations. However, for different brands of olive oil, the peak shapes of the fluorescence intensities and the fluorescence intensity values are quite different, and the degree of change in fluorescence intensity with the oxidation time also varies. This is because the degree of fluorescence inner filter effect caused by different matrices of different brands of olive oil is different. Therefore, it is difficult to accurately predict the PV and AV of different brands of olive oil using only the fluorescence spectrum.

[0060] The absorption coefficient peaks mainly exist in two wavelength ranges: 300 - 500 nm, 650 - 700 nm, mainly μ aThe peaks are around 415, 455, 482, and 670 nm, which are caused by pheophytin a and its derivatives, lutein and its derivatives, β-carotene, pheophytin b, and total chlorophyll, respectively. It can be found that the μ at the fluorescence emission wavelength (375 nm) and the characteristic fluorescence wavelengths (445, 475, 525, 680, 725 nm) of different brands of olive oil a varies greatly. Therefore, the primary and secondary inner filter effect degrees at different characteristic wavelengths in different brands of olive oil are different.

[0061] Determine the candidate characteristic fluorescence wavelengths according to the fluorescence peaks: 445, 475, 525, 680, 725 nm.

[0062] S5. Determination of the optimal characteristic fluorescence wavelength:

[0063] To reflect the influence of the inner filter effect on the fluorescence spectrum prediction of PV and AV, and at the same time reflect the applicability and robustness of the non-destructive determination method of olive oil PV and AV based on the fluorescence inner filter effect correction proposed in the present invention. First, divide the sample set. Divide all samples into a model calibration set and a prediction set, and the samples in the model calibration set and the prediction set cannot overlap. Use the olive oil of 4 brands (brand 1, 2, 3, 5) as the calibration set, and the olive oil of the remaining 1 brand (brand 4) as the verification set;

[0064] Then, respectively, with I. the fluorescence intensity F at the candidate characteristic fluorescence wavelengths (445, 475, 525, 680, 725 nm) em , II. the absorption coefficient μ at the excitation wavelength (375 nm) a,ex , the absorption coefficient μ at the candidate characteristic fluorescence wavelengths (445, 475, 525, 680, 725 nm) a,em and the fluorescence intensity F em , establish a support vector machine regression (SVR) model for the peroxide value and acid value of olive oil, compare the model effects, and the model effects are evaluated by the correlation coefficients (r c , r p ) of the calibration set and the verification set, the root mean square errors (RMSEC, RMSEP). A model with good prediction effect should have higher r c , r p as well as lower RMSEC and RMSEP, and the difference between r c and r p , RMSEC and RMSEP should not be too large, otherwise it is regarded as overfitting or underfitting. The model effects are shown in the following table:

[0065]

[0066] Among them, 445F, 475F, 525F, 680F, and 725F are SVR models established based on the fluorescence intensities at 445, 475, 525, 680, and 725 nm respectively. 445F + 375&445μ a is the SVR model established based on the fluorescence intensity at 445 nm, μ at 375 and 445 nm a ; 475F + 375&475μ a is the SVR model established based on the fluorescence intensity at 475 nm, μ at 375 and 475 nm a ; 525F + 375&525μ a is the SVR model established based on the fluorescence intensity at 525 nm, μ at 375 and 525 nm a ; 680F + 375&680μ a is the SVR model established based on the fluorescence intensity at 680 nm, μ at 375 and 680 nm a ; 725F + 375&725μ a is the SVR model established based on the fluorescence intensity at 725 nm, μ at 375 and 725 nm a . From the results in the table, it can be found that the models of 475F + 375&475μ a have better predictions for both PV and AV. The optimal characteristic fluorescence wavelength is determined to be 475 nm.

[0067] The fluorescence intensities of 5 brands of olive oil at 475 nm, μ at 375 nm and 475 nm a are respectively as Figure 3 , Figure 4 , Figure 5 shown. It can be found that the fluorescence intensity at 475 nm increases significantly with the oxidation time, and this trend is consistent with that of PV and AV. However, due to the significant differences in the initial values and rising speeds of the fluorescence intensities of different brands of olive oil at 475 nm, it is difficult to accurately predict the PV and AV of different brands of olive oil only using the fluorescence intensity. The μ a at 375 nm has no obvious trend with the oxidation time, but the μ a of different brands of olive oil at 375 nm varies greatly, resulting in different primary inner filter effects. The μ a at 475 nm has an obvious downward trend with the oxidation time, and the initial values and downward speeds of different brands of olive oil are significantly different, indicating that the secondary inner filter effects caused by different brands of olive oil at different oxidation times vary greatly.

[0068] S6. Determination of the calculation formulas for peroxide value and acid value based on fluorescence inner filter effect correction:

[0069] According to the results of S5, the optimal characteristic fluorescence wavelength is 475 nm. Therefore, it is necessary to correct the inner filter effect of the fluorescence intensity at 475 nm to achieve accurate prediction of PV and AV of olive oils of different brands.

[0070] According to the fluorescence intensity F of the olive oil samples in the calibration set at 475 nm em , absorption coefficient μ a,em and the absorption coefficient μ at 375 nm a,ex , using non-linear least squares fitting, determine the parameters a, b, c, d, e in the following fluorescence inner filter effect correction formula:

[0071]

[0072] where y are the PV and AV values of the olive oil samples in the calibration set respectively.

[0073] Finally, the calculation formulas for PV and AV are determined as follows:

[0074]

[0075]

[0076] To verify the prediction effects of these two formulas, substitute the fluorescence intensity at 475 nm, absorption coefficient and absorption coefficient at 375 nm of the olive oil samples in the calibration set and validation set into the above formulas for calculation to obtain the calculated values of PV and AV. Plot the distribution diagrams of the calculated values of PV and AV versus the actual values as shown in Figure 6 , Figure 7 . Among them, the effect of the PV calculation formula is: r c = 0.8714, r p = 0.9435, RMSEC = 6.4801, RMSEP = 6.4334; the effect of the PV calculation formula is: r c = 0.8532, r p = 0.9456, RMSEC = 0.1292, RMSEP = 0.1459. The results show that the proposed non-destructive determination method of peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction can achieve the calculation of PV and AV of olive oils of different brands.

[0077] The above are only specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments and there can be many variations. All variations that can be directly derived or associated by those of ordinary skill in the art from the content disclosed in the present invention should be considered within the protection scope of the present invention.

Claims

1. A non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction, characterized in that, It includes the following steps: S1. Prepare a virgin olive oil sample; S2. Measure the peroxide value PV and acid value AV of the virgin olive oil sample, and take the measurement results as the actual values; S3. At a specific excitation wavelength, collect the absorption spectrum and fluorescence spectrum of the sample within a certain fluorescence wavelength range; S4. Spectral analysis: According to the fluorescence spectrum waveforms of virgin olive oil samples of different brands, determine the candidate characteristic fluorescence wavelengths from the positions of the fluorescence peaks; S5. Verify the influence of the fluorescence inner filter effect on the detection results at different candidate characteristic fluorescence wavelengths, and determine the optimal characteristic fluorescence wavelength; Specifically: First, divide the virgin olive oil sample set of different brands. Use one brand of virgin olive oil as the validation set, and the remaining brands of virgin olive oil as the calibration set; Use the fluorescence intensity F at the fluorescence wavelength of the feature to be selected em as the input parameter, and use the peroxide value PV and acid value AV as the output parameters respectively. Use the calibration set to establish the first support vector machine regression model for the peroxide value and acid value of olive oil and, the absorption coefficient μ at the excitation wavelength a,ex , the absorption coefficient μ at the wavelength of the to-be-selected characteristic fluorescence a,em and the fluorescence intensity F em As input parameters, with the peroxide value PV and the acid value AV as output parameters respectively, a second support vector machine regression model for the peroxide value and acid value of olive oil is established using the calibration set; Use the validation set to test the effects of the first support vector machine regression model and the second support vector machine regression model at different candidate characteristic fluorescence wavelengths. Use the correlation coefficient and root mean square error as evaluation indicators, and determine the optimal characteristic fluorescence wavelength according to the test results; S6. According to the fluorescence intensity F at the optimal characteristic fluorescence wavelength em , absorption coefficient μ a,em and the absorption coefficient μ at the excitation wavelength a,ex , using non - linear least - squares fitting, a quantitative prediction model for the peroxide value PV and acid value AV corrected based on the fluorescence inner filter effect is obtained: y1 = F em (a1·10 b1·μa,ex +c1·10 d1·μa,em )+e1 y2 = F em (a2·10 b2·μa,ex +c2·10 d2·μa,em )+e2 Among them, y1 is the peroxide value PV, a1, b1, c1, d1, e1 are the fitting parameters of the quantitative prediction model of the peroxide value PV, y2 is the acid value AV, and a2, b2, c2, d2, e2 are the fitting parameters of the quantitative prediction model of the acid value AV; S6. Collect the fluorescence intensity F at the optimal characteristic fluorescence wavelength of the virgin olive oil to be measured em , absorption coefficient μ a,em and the absorption coefficient μ at the excitation wavelength a,ex . Using the quantitative prediction models of peroxide value PV and acid value AV established in step S6 based on the correction of fluorescence inner filter effect, calculate the peroxide value and acid value of the virgin olive oil to be measured.

2. The non-destructive determination method of peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction according to claim 1, wherein, In the step S1, fill virgin olive oil of different brands into brown bottles, place them in an oven at 110°C for accelerated oxidation treatment, take samples at different time nodes, and cool for later use.

3. The non-destructive determination method for peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction according to claim 1, characterized in that, In the step S3, take a certain amount of olive oil sample, use a tungsten halogen lamp and a laser with a wavelength of 375 nm as light sources respectively, the integration time is 2 s for both, and the smoothing times is 2, to obtain the fluorescence spectrum and absorption spectrum of the olive oil sample in the wavelength range of 200 - 1000 nm.

4. The non-destructive determination method for peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction according to claim 1, characterized in that, The candidate characteristic fluorescence wavelengths include 445 nm, 475 nm, 525 nm, 680 nm, and 725 nm.

5. The non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction according to claim 1, characterized in that, The optimal characteristic fluorescence wavelength is 475 nm.

6. The non-destructive determination method for the peroxide value and acid value of virgin olive oil based on fluorescence inner filter effect correction according to claim 5, wherein, The quantitative prediction models of the peroxide value PV and acid value AV corrected based on the fluorescence inner filter effect are as follows: Among them, F em and μ a,em are the fluorescence intensity and absorption coefficient at 475 nm respectively, and μ a,ex is the absorption coefficient at 375 nm.

Images