CN111103259B - Rapid detection method for frying oil quality based on spectrum technology - Google Patents
Rapid detection method for frying oil quality based on spectrum technology Download PDFInfo
- Publication number
- CN111103259B CN111103259B CN202010091079.9A CN202010091079A CN111103259B CN 111103259 B CN111103259 B CN 111103259B CN 202010091079 A CN202010091079 A CN 202010091079A CN 111103259 B CN111103259 B CN 111103259B
- Authority
- CN
- China
- Prior art keywords
- frying oil
- data
- spectrum
- model
- frying
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001228 spectrum Methods 0.000 title claims abstract description 84
- 238000001514 detection method Methods 0.000 title claims abstract description 34
- 238000005516 engineering process Methods 0.000 title claims abstract description 24
- 238000000034 method Methods 0.000 claims abstract description 55
- 238000012937 correction Methods 0.000 claims abstract description 33
- 230000003595 spectral effect Effects 0.000 claims abstract description 32
- 238000005259 measurement Methods 0.000 claims abstract description 24
- 238000012360 testing method Methods 0.000 claims abstract description 22
- 238000012549 training Methods 0.000 claims abstract description 15
- 239000000523 sample Substances 0.000 claims description 57
- 238000002329 infrared spectrum Methods 0.000 claims description 32
- 239000002253 acid Substances 0.000 claims description 26
- 150000002978 peroxides Chemical class 0.000 claims description 26
- 230000004927 fusion Effects 0.000 claims description 25
- 238000002835 absorbance Methods 0.000 claims description 24
- 238000001237 Raman spectrum Methods 0.000 claims description 19
- 238000007781 pre-processing Methods 0.000 claims description 14
- 230000036961 partial effect Effects 0.000 claims description 12
- 235000013305 food Nutrition 0.000 claims description 11
- 238000011156 evaluation Methods 0.000 claims description 9
- 238000010606 normalization Methods 0.000 claims description 8
- 238000004448 titration Methods 0.000 claims description 6
- 238000001344 confocal Raman microscopy Methods 0.000 claims description 5
- 238000009499 grossing Methods 0.000 claims description 5
- 238000012569 chemometric method Methods 0.000 claims description 4
- 238000010561 standard procedure Methods 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 3
- 239000013307 optical fiber Substances 0.000 claims description 3
- 238000005457 optimization Methods 0.000 claims description 3
- 230000009286 beneficial effect Effects 0.000 claims description 2
- 238000000605 extraction Methods 0.000 claims description 2
- 239000003921 oil Substances 0.000 description 86
- 239000011159 matrix material Substances 0.000 description 14
- 230000000694 effects Effects 0.000 description 11
- 238000002095 near-infrared Raman spectroscopy Methods 0.000 description 9
- 238000004477 FT-NIR spectroscopy Methods 0.000 description 4
- 239000003960 organic solvent Substances 0.000 description 4
- 238000002790 cross-validation Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 238000004445 quantitative analysis Methods 0.000 description 3
- 238000005070 sampling Methods 0.000 description 3
- 239000000126 substance Substances 0.000 description 3
- 239000002699 waste material Substances 0.000 description 3
- 238000002474 experimental method Methods 0.000 description 2
- 238000012417 linear regression Methods 0.000 description 2
- 235000013606 potato chips Nutrition 0.000 description 2
- 238000004497 NIR spectroscopy Methods 0.000 description 1
- 238000001069 Raman spectroscopy Methods 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004440 column chromatography Methods 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 230000002860 competitive effect Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 239000008157 edible vegetable oil Substances 0.000 description 1
- 230000002349 favourable effect Effects 0.000 description 1
- 125000004435 hydrogen atom Chemical group [H]* 0.000 description 1
- 230000007062 hydrolysis Effects 0.000 description 1
- 238000006460 hydrolysis reaction Methods 0.000 description 1
- 238000003703 image analysis method Methods 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000001225 nuclear magnetic resonance method Methods 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 230000003647 oxidation Effects 0.000 description 1
- 238000007254 oxidation reaction Methods 0.000 description 1
- 238000010238 partial least squares regression Methods 0.000 description 1
- 238000006116 polymerization reaction Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000011002 quantification Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000012306 spectroscopic technique Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
- 238000001845 vibrational spectrum Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/3577—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light for analysing liquids, e.g. polluted water
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/62—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light
- G01N21/63—Systems in which the material investigated is excited whereby it emits light or causes a change in wavelength of the incident light optically excited
- G01N21/65—Raman scattering
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Nuclear Medicine, Radiotherapy & Molecular Imaging (AREA)
- Investigating, Analyzing Materials By Fluorescence Or Luminescence (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The method for quickly detecting the quality of the frying oil based on the spectrum technology is characterized by comprising three parts of contents of collecting training set samples and measurement data, establishing a correction model, detecting the correction model by using test set data and detecting certain frying oil, and specifically comprises the following steps of: collecting training set samples and measurement data and establishing a correction model: and (3) detecting the correction model by using the test set data: detecting a certain frying oil: according to the invention, the quality detection of the frying oil is realized by performing spectral scanning on the frying oil sample with unknown components. The method has the advantages of rapidness, no damage, no pollution, convenience and the like, and provides a good foundation for realizing the quality detection of the frying oil in the market.
Description
Technical Field
The invention discloses a quick detection method for frying oil quality based on a spectrum technology, and relates to food safety detection by using the spectrum technology.
Background
With the improvement of living standard, people pay more and more attention to food safety, and fried food is popular in China, so the quality problem of the frying oil becomes more important. The frying oil generates certain harmful substances through oxidation, hydrolysis, polymerization and other reactions in the frying process, and the substances not only influence the quality of the frying oil, but also are harmful to human health. Polar components, peroxide number and acid value in the frying oil are important indexes for measuring the quality of the frying oil.
At present, a plurality of standard measuring methods for polar components, peroxide values and acid values exist, and the national standard measuring method for the polar components is a column chromatography method, but the method has the defects of time and labor waste, large amount of organic solvents, large error and the like; the national standard determination method for peroxide value and acid value is a titration method, which has the defects of needing a large amount of organic solvent and damaging samples, and other detection methods, such as a nuclear magnetic resonance method, an image analysis method, an electronic nose and a conductivity method, are not widely used due to the high cost of instruments, complicated operation steps, the need of using a large amount of organic solvent and the like. In recent years, the spectroscopic technology is rapidly developed and widely applied due to the unique advantages of rapid nondestructive detection, and particularly, the spectroscopic technology is fully determined by the unique advantages of simple operation, no need of consuming organic solvents and the like in the field of food safety detection. However, in the prior art, the quality of the food is not detected by the spectroscopic technique.
Disclosure of Invention
The invention provides a quick detection method for frying oil quality based on a spectrum technology, which overcomes the defects that the detection of frying oil needs to directly measure the polar component value, the peroxide value or the acid value, so that the manual operation time is long and the detection result is slow.
The method for quickly detecting the quality of the frying oil based on the spectrum technology is characterized by comprising three parts of contents of collecting training set samples and measurement data, establishing a correction model, detecting the correction model by using test set data and detecting certain frying oil, and specifically comprises the following steps of:
collecting training set samples and measurement data and establishing a correction model:
frying food with frying oil, frying the food with the same amount in each unit time, extracting frying oil samples according to a certain time interval, numbering each sample, and recording the frying time of each numbered sample;
step two, detecting and recording the polar component value, the peroxide value and the acid value of each sample by adopting a national standard method;
step three, scanning and recording various kinds of spectral data of each sample;
respectively carrying out corresponding pretreatment on the spectrum of each sample, and carrying out normalization treatment to obtain a normalization data form;
fifthly, extracting characteristic wave bands of the normalized form of the spectrum data of all the samples obtained in the fourth step in various spectrums, and then fusing the spectrum data of a plurality of characteristic wave bands in the various extracted spectrums;
step six, aiming at the polar component values, peroxide values or acid values of the frying oil samples with different frying durations in the training set samples, establishing a correction model by combining the data fused in the step five with a chemometric method;
and (3) detecting the correction model by using the test set data:
collecting a test set sample and measurement data; collecting the test set data samples and the measurement data in the same steps from one step to the third step, and predicting the components of the test set samples by using the correction model obtained in the sixth step; component prediction refers to predicting the polar component value, peroxide value or acid value;
step eight, evaluating the correction model by adopting model evaluation parameters; if the evaluation is unqualified, optimizing the model; after optimization, turning to the seventh step until the model is qualified;
if the model is qualified, the model is taken as a selected model;
detecting a certain frying oil:
performing spectrum scanning related to a selected model on a frying oil sample to be detected, and then predicting component values by using the selected model to obtain corresponding component values, wherein the component values refer to polar component values, peroxide values or acid values;
and step ten, comparing the predicted corresponding component values with internationally corresponding component values to determine the quality of the frying oil to be detected.
Wherein the preprocessing described in step four is data preprocessing of the spectral data of the frying oil sample by a smoothing method, a baseline correction or a derivative preprocessing method.
And thirdly, obtaining absorbance data of the corresponding spectrum, wherein the absorbance data of the corresponding spectrum obtained in the third step are near infrared spectrum absorbance data, intermediate infrared spectrum absorbance data and Raman spectrum absorbance data.
The step five of extracting the characteristic bands from the plurality of spectra refers to extracting the characteristic bands from two or more spectra, and selecting the extracted characteristic bands in each spectrum by using a correlation coefficient method or an interval partial least square method.
And in the fifth step, the spectral data fusion is performed on the plurality of characteristic bands in all the extracted plurality of spectra, which means that one or more characteristic bands are obtained in each of the plurality of spectra, and the spectral data fusion is performed on all the characteristic bands obtained in the plurality of spectra.
The acquisition method of the near infrared spectrum is a transmission type optical fiber probe, the detection amount of the frying oil sample is bottled, and shaking measurement is adopted;
the collection mode of the mid-infrared spectrum is diffuse reflection type, and a pipette is adopted for quantitative titration;
wherein, the collection instrument of the Raman spectrum is a DXR laser confocal micro-Raman spectrometer.
Compared with the prior art, the invention has the beneficial effects that:
according to the invention, the quality detection of the frying oil is realized by performing spectral scanning on the frying oil sample with unknown components. The method has the advantages of rapidness, no damage, no pollution, convenience and the like, and provides a good foundation for realizing the quality detection of the frying oil in the market. The quality of the frying oil can be rapidly and nondestructively detected. The invention combines the advantages of different spectrum technologies, complements the advantages of various spectra, improves the accuracy of the detection model, and expands the application range of the model, and the comparison with the model effect of a single spectrum technology proves that the accuracy of predicting the content of the unknown components in the sample by the model established by combining the multisource spectral feature fusion and the partial least square method can meet the requirement of the frying oil, so the method for detecting the quality of the frying oil by the multisource spectral fusion technology can well realize the quick, nondestructive and pollution-free detection of the quality of the frying oil.
Drawings
FIG. 1 is a flow chart of a multi-source spectral data fusion modeling analysis of the present invention;
FIG. 2 shows a frying pan of the present invention during a frying experiment;
FIG. 3 is a diagram of the experimental process of frying in accordance with the present invention;
FIG. 4 (A) is a graph of the effect of a prior art near infrared spectrum on the quantitative modeling of the polar component values of frying oil;
FIG. 4 (B) is a predictive graph of a prior art near infrared spectroscopy vs. frying oil polar component value quantitative model vs. polar component values for test set samples (comparative);
FIG. 5 (A) is a graph showing the effect of a model for quantitatively modeling the polar component values of frying oil by using a Raman spectrum according to the prior art, and FIG. 5 (B) is a graph showing the effect of a model for quantitatively modeling the polar component values of frying oil by using a Raman spectrum according to the prior art on the polar component values of frying oil according to the prior art
Mapping (for comparison);
FIG. 6 (A) is a graph of the effect of a model for the near infrared-Raman spectroscopy full-spectral cepstrum data fusion on the quantitative modeling of the polar component values of frying oil;
FIG. 6 (B) is a prediction chart of polar component values of a test set sample of a quantitative model of polar component values of frying oil by near infrared-Raman spectroscopy full-wave spectral data fusion (simple data fusion effect, for comparison) according to the present invention;
FIG. 7 (A) is a graph of the effect of the near infrared-Raman spectrum feature fusion model on the quantitative modeling of the polar component values of frying oil according to the present invention;
FIG. 7 (B) is a prediction graph of the near infrared-Raman spectrum feature fusion of the present invention on the polar component values of the frying oil polar component value quantitative model versus the polar component values of the test set samples (optimal model effect graph, used for the conclusion, also for the effect demonstration of the patent application);
fig. 8 is a table of comparing performance indexes based on different model effect diagrams corresponding to fig. 4, 5, 6, and 7, respectively.
Detailed Description
The method for quickly detecting the quality of the frying oil based on the spectrum technology is characterized by comprising three parts of contents of collecting training set samples and measurement data, establishing a correction model, detecting the correction model by using test set data and detecting certain frying oil, and specifically comprises the following steps of:
collecting training set samples and measurement data and establishing a correction model:
frying food with frying oil, frying the food with the same amount in each unit time, extracting frying oil samples according to a certain time interval, numbering each sample, and recording the frying time of each numbered sample;
detecting and recording the polar component value, the peroxide value and the acid value of each sample by adopting a national standard method;
step three, scanning and recording various kinds of spectral data of each sample;
respectively carrying out corresponding pretreatment on the spectrum of each sample, and carrying out normalization treatment to obtain a normalization data form;
fifthly, extracting characteristic wave bands of the normalized form of the spectrum data of all the samples obtained in the step four in various spectrums, and then fusing the spectrum data of a plurality of characteristic wave bands in the extracted various spectrums;
step six, aiming at the polar component values, peroxide values or acid values of the frying oil samples with different frying durations in the training set samples, establishing a correction model by combining the data fused in the step five with a chemometric method;
and (3) detecting the correction model by using the test set data:
collecting a test set sample and measurement data; collecting the test set data samples and the measurement data in the same steps from one step to the third step, and predicting the components of the test set samples by using the correction model obtained in the sixth step; component prediction refers to predicting the polar component value, peroxide value or acid value;
step eight, evaluating the correction model by adopting model evaluation parameters; if the evaluation is unqualified, optimizing the model; after optimization, turning to the seventh step until the model is qualified;
if the model is qualified, the model is taken as a selected model;
detecting a certain frying oil:
performing spectrum scanning related to a selected model on a frying oil sample to be detected, and then predicting component values by using the selected model to obtain corresponding component values, wherein the component values refer to polar component values, peroxide values or acid values;
and step ten, comparing the corresponding component values obtained by prediction with the internationally corresponding component values to determine the quality of the frying oil to be detected.
Wherein the preprocessing described in step four is data preprocessing of the spectral data of the frying oil sample by a smoothing method, a baseline correction or a derivative preprocessing method.
And thirdly, obtaining absorbance data of the corresponding spectrum, wherein the absorbance data of the corresponding spectrum obtained in the third step are near infrared spectrum absorbance data, intermediate infrared spectrum absorbance data and Raman spectrum absorbance data.
The step five of extracting the characteristic bands from the plurality of spectra refers to extracting the characteristic bands from two or more spectra, and selecting the extracted characteristic bands in each spectrum by using a correlation coefficient method or an interval partial least square method.
And in the fifth step, the spectral data fusion is performed on the plurality of characteristic bands in all the extracted plurality of spectra, which means that one or more characteristic bands are obtained in each of the plurality of spectra, and the spectral data fusion is performed on all the characteristic bands obtained in the plurality of spectra.
The acquisition method of the near infrared spectrum is a transmission type optical fiber probe, the detection amount of the frying oil sample is bottled, and shaking measurement is adopted;
the collection mode of the mid-infrared spectrum is diffuse reflection type, and a pipette is adopted for quantitative titration;
wherein, the collection instrument of the Raman spectrum is a DXR laser confocal micro-Raman spectrometer.
The technical scheme provided by the invention is as follows:
the method for quickly detecting the quality of the frying oil based on the spectrum technology is characterized in that different spectrum data of a frying oil sample are fused under a specific method, and a quantitative model based on the polar component value, the peroxide value or the acid value of the frying oil is established by combining characteristic information which is favorable for modeling in two spectra, so that the method for quickly and nondestructively detecting the quality of the frying oil is realized, and the method comprises the following specific steps:
1) Collecting frying oil samples of different frying time of the frying oil, and determining a training set sample and a testing set sample;
2) Detecting the polar component value, the peroxide value or the acid value of the frying oil samples with different frying time lengths by adopting a national standard detection method;
2) Collecting various spectral data of frying oil samples with different frying time lengths;
2) Respectively carrying out corresponding pretreatment on the collected spectral data to obtain a normalized data form;
2) Carrying out feature extraction on the preprocessed and normalized different spectral data obtained in the step 4), and then carrying out fusion on the spectral data;
2) Aiming at the polar component values, peroxide values or acid values of the frying oil samples with different frying durations in the training set, establishing a correction model for the data fused in the step 5) by using a chemometric method;
2) Predicting components (polar component values or peroxide values or acid values) of the samples in the test set by using various spectrum fusion data by using the correction model obtained in the step 6) according to the polar component values, the peroxide values or the acid values of the frying oil samples with different frying durations;
2) And evaluating the correction model by adopting model evaluation parameters.
Aiming at the quality detection method of the frying oil of the multi-source spectrum fusion technology, further,
the frying conditions of the frying oil sample in the step 1) are as follows: the frying temperature is constant at 180 ℃, the fried object is the chips, 1kg of chips are fried twice per hour, the sampling is carried out once per two hours on average, until the polar component value or the peroxide value or the acid value in the frying oil exceeds the national detection standard, the chips are waste oil, and the frying sampling is not carried out.
Measuring the polar component value, peroxide value and acid value of the frying oil sample in the step 2) by a national standard method, and referring to GB/T5009.202-2016 for the specific characteristics of corresponding standards
The spectral data required to be collected in the step 3) of collecting the frying oil sample comprise near infrared spectrum absorbance data, intermediate infrared spectrum absorbance data, raman spectrum absorbance data and the like.
The acquisition instrument of the near infrared spectrum and the mid infrared spectrum comprises: an Antaris II Fourier transform near Infrared Spectroscopy (FT-NIR) from Saimer Feishale, having a wavelength range of: 3800-12000cm -1
The near infrared spectrum measurement mode of the frying oil is diffuse reflection of a light ray probe, the detection amount of a frying oil sample is 60 milliliters in a bottle, and shaking measurement is adopted.
The collection mode of the mid-infrared spectrum is diffuse reflection type, and a pipette is adopted for quantitative titration.
The acquisition instrument of the Raman spectrum is as follows: DXR laser confocal micro-Raman spectrometer, laser light source: 780nm, wave number range: 50-3500cm -1
The spectrum preprocessing in the step 4) is to perform data preprocessing on the spectrum data of the frying oil sample by using preprocessing methods such as a smoothing method, baseline correction or derivative method and the like. And after the data preprocessing and normalization are finished, respectively extracting the characteristics of different spectral data. The invention adopts a correlation coefficient method and an interval partial least square method (IPLS) to extract the characteristics of the two preprocessed spectral absorbance data.
The correlation coefficient method is to perform correlation calculation on the absorbance vector corresponding to each wavelength in the training set sample spectral array and the concentration of the component to be measured in the concentration array to obtain a wavelength-correlation coefficient R diagram (or R 2 Graph) the more information should be given for wavelengths for which the absolute value of the correlation coefficient (or the decision coefficient) is larger.
The principle of the method is that the whole spectrum is equally divided into a plurality of sub-intervals with equal width, PLS regression is carried out on each sub-interval, the interval corresponding to the minimum RMSECV is found out, and then the wavelength variable is expanded in a single way or in two ways by taking the interval as the center, so that the optimal wavelength interval is obtained.
In the step 6), a model is established for the polar component values, peroxide values or acid values in the frying oil with different frying time lengths by using a partial least square method, and the principle of partial least square is as follows:
first, a spectrum matrix X and a concentration matrix Y are decomposed according to formula 1 and formula 2:
x = TP + E (formula 1)
Y = UQ + F (formula 2)
Wherein T and U are respectively the scoring matrix of X and Y; p and Q are the load matrix of X and Y respectively; e and F are fitting residual matrixes of PLS of X and Y respectively;
and secondly, performing linear regression of T and U matrixes according to the formulas 3 and 4:
u = TB (formula 3)
B＝(T T T) -1 T T Y (formula 4)
When quantitative analysis is carried out on the prediction sample, the spectrum matrix Y of the sample to be measured is obtained according to the matrix P in the formula 1 Is unknown Score matrix T of Is unknown To obtain a predicted value of the concentration of the component according to equation 5:
Y is unknown ＝T Is unknown BQ (formula 5)
Thus, the polar component value, the peroxide value and the acid value of the frying oil can be quantified.
Extracting characteristic bands
When chemometrics school is performed, it is not necessary that all spectral data participate in calibration, and a good calibration effect can be obtained by selecting certain spectral regions, so that the spectral regions need to be selected. Possible methods for selecting the variables are: correlation coefficient methods, competitive adaptive re-weighted sampling (CARS), and interval and moving window partial least squares (IPLS), but are not limited to the above methods.
The established PLS algorithm quantitative analysis model mainly has the following two evaluation indexes:
(1) Determining the coefficient (R) 2 ): for calculating the sample pass spectrumCorrelation between the predicted values of the method and the actual measured values obtained by conventional detection methods. Under the premise of same concentration range, R 2 The closer to 1, the better the correlation between the predicted result and the true value, R 2 Calculated according to equations 2-6:
in the formula (I), the compound is shown in the specification,
and y i Represents the prediction and the true measurement, respectively, of the ith sample>
Representing the average of the true measurements of all samples.
(2) Cross validation Root Mean Square Error (RMSECV): the index is calculated by using an interactive verification method and is used for evaluating whether the model building is feasible and quantifying the prediction capability of the model. RMSECV is calculated according to equations 2-7:
in the formula, y i Represents the chemical truth value of the ith sample,
representing the result of the model measurement sample i after the ith sample is removed from the sample set, and n represents the number of corrected lumped samples.
The invention will be further described by way of examples, without in any way limiting the scope of the invention, with reference to the accompanying drawings.
The invention provides a quick detection method for frying oil quality based on a spectrum technology, which can realize quick and nondestructive detection of frying oil quality. The invention combines the advantages of different spectrum technologies, complements the advantages of various spectra, improves the accuracy of the detection model, expands the application range of the model, and can well realize the quick, nondestructive and pollution-free detection of the quality of the frying oil.
Example (b):
1. experimental Material
The frying oil used in the experiment is novel frying oil newly developed by grain scientific research institute in Beijing, the capacity of a frying pan is 20 liters, the fried object is chips, and the frying conditions are as follows: keeping the temperature at 180 ℃, frying the potato chips twice per hour for 1kg, continuously frying the potato chips once every two hours, simultaneously measuring the value of the polar components in the frying oil until the value of the polar components in the frying oil is 27%, stopping frying, and marking the frying oil as waste oil. The polar component value in the frying oil is measured by a Testo270 edible oil quality detector according to GB/T500.202-2016.
2. Spectrum collection
The spectral data required to be collected of the frying oil sample comprise near infrared spectrum absorbance data, intermediate infrared spectrum absorbance data, raman spectrum absorbance data and the like.
The acquisition instrument of the near infrared spectrum and the mid infrared spectrum comprises: an Antaris II Fourier transform near Infrared Spectroscopy (FT-NIR) from Saimer Feishale, having a wavelength range of: 3800-12000cm-1
The near infrared spectrum measurement mode of the frying oil is diffuse reflection of a light ray probe, the detection amount of a frying oil sample is 60 milliliters in a bottle, and shaking measurement is adopted.
The collection mode of the mid-infrared spectrum is diffuse reflection type, and a pipette is adopted for quantitative titration.
The acquisition instrument of the Raman spectrum is as follows: DXR laser confocal micro-Raman spectrometer, laser light source: 780nm, wave number range: 50-3500cm-1 respectively carries out corresponding pretreatment on the collected near infrared spectrum and Raman spectrum data, wherein the spectrum pretreatment specifically comprises obtaining a normalized data form by a smoothing method, baseline correction or derivative and the like.
3. And after the data preprocessing and normalization are finished, respectively extracting the characteristics of different spectral data. The invention adopts a correlation coefficient method and an interval partial least square method (IPLS) to extract the characteristics of the two preprocessed spectral data. Obtaining a near infrared spectrum characteristic wave band: 4798-5597, 5600-6395, and 6400-7220, to obtain Raman spectrum characteristic bands 80-401, 410-729, and 1046-1690, and performing serial fusion of near infrared-Raman spectrum after obtaining the above characteristic spectra.
4. Model building
After the near infrared-Raman spectrum is fused, a model is established for polar component values in the frying oil with different frying time lengths by utilizing a partial least square method, wherein the partial least square principle is as follows:
first, a spectrum matrix X and a concentration matrix Y are decomposed according to formula 1 and formula 2:
x = TP + E (formula 1)
Y = UQ + F (formula 2)
Wherein T and U are respectively the scoring matrix of X and Y; p and Q are the load matrix of X and Y respectively; e and F are fitting residual matrixes of PLS of X and Y respectively;
secondly, linear regression of the T and U matrices is performed according to equations 3 and 4:
u = TB (formula 3)
B＝(T T T) -1 T T Y (formula 4)
When quantitative analysis is carried out on the prediction sample, the spectrum matrix Y of the sample to be measured is obtained according to the matrix P in the formula 1 Is unknown Score matrix T of Is unknown To obtain a predicted value of the concentration of the component according to equation 5:
Y is unknown ＝T Is unknown BQ (formula 5)
Thereby achieving quantification of the polar component value of the frying oil;
and (3) analyzing a modeling result:
after the obtained near infrared-raman spectral feature fusion data is combined, a quantitative model is established for a training set sample by combining a partial least square method, a modeling result is shown in fig. 7 (a), a prediction set result is shown in fig. 7 (B), a model evaluation result is that a decision coefficient is 0.9680, a cross validation root mean square error is 0.5696, the decision coefficient is improved by 66% compared with a result obtained by singly modeling near infrared spectral data of frying oil (shown in fig. 4) and singly modeling raman spectral data (shown in fig. 5), the cross validation root mean square error is improved by 32%, and performance indexes are compared and shown in table 1. The comparison graph comprises a quantitative model of the polar component values in the frying oil by near infrared-Raman spectrum data level fusion, the near infrared-Raman spectrum data level fusion utilizes a full-wave common weighted average fusion method, and after comparison, multispectral feature fusion is expected to obviously improve the accuracy of modeling of the polar component values of the frying oil and enhance the robustness.
The invention particularly relates to the realization of a near infrared spectrum and Raman spectrum data fusion technology and a near infrared spectrum and Raman spectrum characteristic fusion technology.
The multi-sensor information fusion is a multi-stage and multi-level data processing process, and mainly comprises 3 levels: the data layer, the characteristic layer and the decision layer have different advantages and disadvantages, and the multi-sensor information fusion system has stronger robustness due to redundancy and complementarity of multi-sensor information.
In addition, there are many spectrum detection methods, such as infrared spectrum detection technology, raman spectrum detection technology, terahertz technology, etc., different spectrum detection technologies all have their unique advantages and disadvantages, for example, near infrared spectrum and raman spectrum all belong to the family of vibration spectrum, the absorption region of the resultant frequency and the frequency multiplication of the vibration of hydrogen group (C-H, O-H) in frying oil is consistent with the near infrared spectrum region, meanwhile, "-C = C —" in frying oil, etc. has great contribution to the molecular vibration in raman spectrum, and the two spectra have complementarity and redundancy.
Claims (1)
1. The method for quickly detecting the quality of the frying oil based on the spectrum technology is characterized by comprising three parts of contents of collecting training set samples and measurement data, establishing a correction model, detecting the correction model by using test set data and detecting certain frying oil, and specifically comprises the following steps of:
collecting training set samples and measurement data and establishing a correction model:
frying food with frying oil, frying the food with the same amount in each unit time, extracting frying oil samples according to a certain time interval, numbering each sample, and recording the frying time of each numbered sample;
detecting and recording the polar component value, the peroxide value and the acid value of each sample by adopting a national standard method;
scanning various spectrum data of each sample to obtain absorbance data of a corresponding spectrum, and recording the absorbance data; the absorbance data refers to near infrared spectrum absorbance data, intermediate infrared spectrum absorbance data and Raman spectrum absorbance data; the acquisition method of the near infrared spectrum is a transmission type optical fiber probe, the detection amount of the frying oil sample is bottled, and shaking measurement is adopted; the collection mode of the mid-infrared spectrum is diffuse reflection type, and a pipette is adopted for quantitative titration; wherein the collection instrument of the Raman spectrum is a DXR laser confocal micro-Raman spectrometer;
respectively carrying out corresponding pretreatment on the spectrum of each sample, and carrying out normalization treatment to obtain a normalization data form; the preprocessing in the step is to carry out data preprocessing on the spectral data of the frying oil sample by a smoothing method, a baseline correction method or a derivative preprocessing method;
fifthly, extracting characteristic wave bands of the normalized form of the spectrum data of all the samples obtained in the step four in various spectrums, and then fusing the spectrum data of a plurality of characteristic wave bands in the extracted various spectrums; the characteristic wave band extraction in the plurality of spectra is to extract the characteristic wave bands in two or more spectra, and the extracted characteristic wave bands are selected in each spectrum by using a correlation coefficient method or an interval partial least square method; after one or more characteristic wave bands are obtained in each spectrum in the multiple spectrums, performing spectrum data fusion on all the characteristic wave bands obtained in the multiple spectrums; by combining characteristic information which is beneficial to modeling in the two spectra, a quantitative model based on the polar component value, the peroxide value or the acid value of the frying oil is further established, so that the quality detection of the frying oil without damage is realized;
step six, aiming at the polar component values, peroxide values or acid values of the frying oil samples with different frying durations in the training set samples, establishing a correction model by combining the spectrum data fused in the step five with a chemometric method;
and (3) detecting the correction model by using the test set data:
collecting a test set sample and measurement data; collecting the test set data samples and the measurement data in the same steps from one step to the third step, and predicting the components of the test set samples by using the correction model obtained in the sixth step; component prediction refers to predicting the polar component value, peroxide value or acid value;
step eight, evaluating the correction model by adopting model evaluation parameters; if the evaluation is unqualified, optimizing the model; after optimization, turning to the seventh step until the model is qualified;
if the model is qualified, the model is taken as a selected model;
detecting a certain frying oil:
performing spectrum scanning related to a selected model on a frying oil sample to be detected, and then predicting component values by using the selected model to obtain corresponding component values, wherein the component values refer to polar component values, peroxide values or acid values;
and step ten, comparing the predicted corresponding component values with the corresponding component values of national standards to determine the quality of the frying oil to be measured.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010091079.9A CN111103259B (en) | 2020-02-13 | 2020-02-13 | Rapid detection method for frying oil quality based on spectrum technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010091079.9A CN111103259B (en) | 2020-02-13 | 2020-02-13 | Rapid detection method for frying oil quality based on spectrum technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111103259A CN111103259A (en) | 2020-05-05 |
| CN111103259B true CN111103259B (en) | 2023-04-07 |
Family
ID=70427916
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010091079.9A Active CN111103259B (en) | 2020-02-13 | 2020-02-13 | Rapid detection method for frying oil quality based on spectrum technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111103259B (en) |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104807775B (en) * | 2015-01-29 | 2019-11-12 | 湖南省农产品加工研究所 | It is a kind of for identify frying oil quality NIR spectra analysis model and method |
| WO2017002079A1 (en) * | 2015-06-30 | 2017-01-05 | Ambifood, Lda | Device and method for measuring the quality of frying oil |
| JP2018205226A (en) * | 2017-06-08 | 2018-12-27 | 昭和産業株式会社 | Analysis of frying oil composition using near infrared spectroscopy |
| CN107894408B (en) * | 2017-11-24 | 2021-04-13 | 中国农业科学院油料作物研究所 | Edible oil multi-element adulteration identification method based on near-infrared spectrometer |
| CN108362661A (en) * | 2017-12-22 | 2018-08-03 | 无锡中科恒源信息科技有限公司 | Frying oil on-line detecting system and its method based on spectrum sensing technology |
| CN110672578A (en) * | 2019-09-25 | 2020-01-10 | 北京市粮食科学研究院 | Model universality and stability verification method for polar component detection of frying oil |
-
2020
- 2020-02-13 CN CN202010091079.9A patent/CN111103259B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN111103259A (en) | 2020-05-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fan et al. | Prediction of soluble solids content and firmness of pears using hyperspectral reflectance imaging | |
| CN1982872B (en) | Near-infrared diffuse reflection spectral method for fastly inspecting drop water content | |
| US20100297291A1 (en) | Visible/near-infrared spectrum analyzing method and grape fermenting method | |
| Xue et al. | Application of particle swarm optimization (PSO) algorithm to determine dichlorvos residue on the surface of navel orange with Vis-NIR spectroscopy | |
| CN105424641B (en) | A kind of near infrared spectrum recognition methods of crude oil type | |
| CN107703097B (en) | Method for constructing model for rapidly predicting crude oil property by using near-infrared spectrometer | |
| Xu et al. | Factors influencing near infrared spectroscopy analysis of agro-products: a review | |
| Fernández-Novales et al. | A feasibility study on the use of a miniature fiber optic NIR spectrometer for the prediction of volumic mass and reducing sugars in white wine fermentations | |
| Li et al. | Nondestructive detection of frying times for soybean oil by NIR-spectroscopy technology with Adaboost-SVM (RBF) | |
| CN111999258B (en) | Spectral baseline correction-oriented weighting modeling local optimization method | |
| CN112179871A (en) | Method for nondestructive detection of caprolactam content in sauce food | |
| CN107966420B (en) | Method for predicting crude oil property by near infrared spectrum | |
| CN110672578A (en) | Model universality and stability verification method for polar component detection of frying oil | |
| CN109100315B (en) | Wavelength selection method based on noise-signal ratio | |
| CN101788459B (en) | Quasi-continuous spectroscopic wavelength combination method | |
| CN102854151B (en) | Chemometrics method for classifying sample sets in spectrum analysis | |
| CN114414521A (en) | Milk main component measuring method based on infrared multispectral sensor | |
| Zhou et al. | Applications of near infrared spectroscopy in cotton impurity and fiber quality detection: A review | |
| CN113484278A (en) | Tomato comprehensive quality nondestructive testing method based on spectrum and principal component analysis | |
| CN110887798B (en) | Nonlinear full-spectrum water turbidity quantitative analysis method based on extreme random tree | |
| CN110865046B (en) | Method for rapidly detecting content of trans-fatty acid isomer of edible oil | |
| CN111103259B (en) | Rapid detection method for frying oil quality based on spectrum technology | |
| CN107271389A (en) | A kind of spectral signature variable fast matching method based on index extreme value | |
| Ning et al. | Study on apple damage detecting method based on relaxation single-wavelength laser and convolutional neural network | |
| CN111896497A (en) | Spectral data correction method based on predicted value |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |