CN109406419B - Method for predicting content of p-hydroxybenzoic acid in wolfberry based on hyperspectral imaging technology - Google Patents
Method for predicting content of p-hydroxybenzoic acid in wolfberry based on hyperspectral imaging technology Download PDFInfo
- Publication number
- CN109406419B CN109406419B CN201811285081.9A CN201811285081A CN109406419B CN 109406419 B CN109406419 B CN 109406419B CN 201811285081 A CN201811285081 A CN 201811285081A CN 109406419 B CN109406419 B CN 109406419B
- Authority
- CN
- China
- Prior art keywords
- sample
- content
- data
- hydroxybenzoic acid
- hyperspectral imaging
- 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
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
Abstract
The invention discloses a model establishing method and a content measuring method for predicting the content of p-hydroxybenzoic acid in wolfberry based on a hyperspectral imaging technology. The method comprises the following steps: 1) performing spectrum scanning on a sample by using a hyperspectral imaging system, and collecting hyperspectral data of 400-1000nm and 1000-2400 nm; 2) sequentially performing RAD correction on the original hyperspectral data, converting the original hyperspectral data into relative reflectivity data by adopting a flat field processing method, processing the data by adopting a multivariate scattering correction method, and analyzing and reducing the dimension of a principal component; performing principal component analysis on the effective waveband to obtain main spectral information; 3) modeling the main spectral information and the content of the p-hydroxybenzoic acid by using principal component regression to obtain a prediction model of the content of the p-hydroxybenzoic acid in the medlar; 4) processing the sample to be detected according to the steps 1) -2), inputting the obtained main spectral information into a prediction model, and calculating to obtain the content of the p-hydroxybenzoic acid in the sample to be detected.
Description
Technical Field
The invention belongs to the field of quality identification of traditional Chinese medicinal materials, and particularly relates to a method for predicting the content of p-hydroxybenzoic acid in wolfberry based on a hyperspectral imaging technology.
Background
The medlar resource is widely distributed and various in China, according to the description of Chinese plant journal, medlar plants have 7 varieties of 3 varieties in China, and northern China, such as northern Hebei, inner Mongolia, northern Shanxi, northern Shaanxi, Gansu, Ningxia, Qinghai and Xinjiang, have wild plants in China, and the fruits are gradually cultivated as the medicines, so that cultivation is mainly performed. According to the examination of many aspects, Ningxia is now used to locate the genuine producing area of the medicinal wolfberry fruit. However, the quality of wolfberry has different levels due to the abundant varieties, more production places, and the influence of various environmental factors such as temperature, precipitation, sunshine and the like in various regions, which is particularly expressed as the difference of the content components of wolfberry. In the market circulation, methods for identifying the chemical components and traditional chemical component contents are mostly based on experience. The error of empirical identification is large, the subjectivity is strong, and the operation method of chemical inspection is complex, time-consuming and labor-consuming.
In recent years, the hyperspectral imaging technology has been developed rapidly, and is applied to the field of aerospace at the earliest. And performing geological exploration and ore identification. Then, the method steps into the agricultural field, and the quality and the type of the agricultural crops are evaluated and distinguished. Therefore, the technical innovation of introducing the hyperspectral imaging technology into the field of traditional Chinese medicines for traditional Chinese medicine identification becomes possible.
Disclosure of Invention
The invention aims to provide a model building method for predicting the content of p-hydroxybenzoic acid in wolfberry based on a hyperspectral imaging technology.
The model establishing method for predicting the content of p-hydroxybenzoic acid in the wolfberry fruit based on the hyperspectral imaging technology comprises the following steps:
1) establishing a sample spectrum:
collecting dried fructus Lycii products of different varieties and production places as sample set; performing spectrum scanning on the samples in the sample set by using a hyperspectral imaging system, and collecting 400-1000nm and 1000-2400nm hyperspectral data to obtain sample set spectra;
2) sample spectrum pretreatment:
a1) RAD (Radiometric calibration) correction is carried out on the original hyperspectral data of the sample;
b1) converting the RAD corrected data into relative reflectivity data by adopting a flat field processing method;
c1) processing the relative reflectivity data by a multivariate scattering correction method;
d1) performing principal component analysis and dimensionality reduction on the relative reflectivity data subjected to the multivariate scattering correction, performing principal component regression analysis by using factors subjected to dimensionality reduction, and selecting a factor number corresponding to the first occurrence of numerical stability and invariability when a regression equation R and an R square are more than 0.99 as an optimal factor number;
e1) performing correlation analysis on the relative reflectivity data after the multivariate scattering correction and the content of the p-hydroxybenzoic acid in the sample set measured by a chemical method, and screening out a wave band with a correlation number of more than 0.4 and a significant level reached by a significance test t test to determine the wave band as an effective wave band;
f1) based on the most appropriate factor number screened out by d1), carrying out principal component analysis on the effective wave band to remove redundant spectral information, and obtaining main spectral information;
3) establishing a correction model: and modeling the main spectral information obtained by the sample set spectrum and the p-hydroxybenzoic acid content measured by a chemical method by using principal component regression to obtain a prediction model of the p-hydroxybenzoic acid content in the wolfberry.
In step 1) of the method, the number of samples in the sample set is greater than or equal to 100.
In the step 1) of the method, the hyperspectral imaging system is specifically a HySpex series hyperspectral imaging spectrometer.
The conditions for the spectral scan are as follows: the distance between a lens of the hyperspectral imager and the wolfberry fruit is 20-30cm, and the moving speed of a platform is 1.5 mm/s; the integration time was 4350 μ s when the collected spectral range was 400-1000nm, and the frame time was 18000; the integration time was 4500 μ s and the frame time was 46928 when the collected spectral range was 1000-2400 nm.
In the step 2), selecting an interested region of the wolfberry fruit by utilizing ENVI; and exporting the average spectral value extracted from the region of interest, preprocessing the initial relative reflectance value, screening a preprocessing method, and finally determining the initial relative reflectance value as the multivariate scattering correction.
In step 2) of the above method, the number of the suitable factors is finally determined to be 10.
In step 2) of the method, the effective waveband determined in the step e1) is 28 to 108 wavebands (the wavelength range is 553 +/-3 nm to 987 +/-3 nm).
In the step 2) of the method, SPSS software is used for analyzing the main components.
In the step 3) of the method, the chemical method for measuring the content of p-hydroxybenzoic acid in the sample set is a liquid-mass spectrometry.
In the step 3) of the method, matlab software is adopted to establish the principal component regression model.
The invention also aims to provide a method for predicting the content of p-hydroxybenzoic acid in the wolfberry fruit based on the hyperspectral imaging technology.
The invention provides a method for predicting the content of p-hydroxybenzoic acid in wolfberry based on a hyperspectral imaging technology, which comprises the following steps:
A) establishing a spectrum of a sample to be detected:
performing spectrum scanning on a sample to be detected by using a hyperspectral imaging system, and collecting 400-1000nm and 1000-2400nm hyperspectral data to obtain a sample set spectrum to be detected;
B) preprocessing the spectrum of a sample to be detected:
a2) performing radiation calibration on original hyperspectral data of a sample;
b2) converting the data after radiation correction into relative reflectivity data by adopting a flat field processing method;
c2) processing the relative reflectivity data by a multivariate scattering correction method;
d2) screening the spectral information of the required effective waveband from the data processed in the step c2) according to the effective waveband determined in the step e 1); performing principal component analysis on the effective waveband to remove redundant spectral information to obtain main spectral information of the sample to be detected;
C) and inputting the main spectral information of the sample to be detected into a prediction model of the p-hydroxybenzoic acid content in the wolfberry fruit, and calculating to obtain the p-hydroxybenzoic acid content in the sample to be detected.
The invention has the following advantages:
the hyperspectral imaging technology is adopted, and the hyperspectral imaging technology is applied to the field of traditional Chinese medicine content prediction, so that the market circulation monitoring of genuine medicinal materials is facilitated; the cost of manual identification is reduced, and the accuracy and the scientificity of identification are improved. The invention applies hyperspectrum to Chinese medicinal material content prediction, and the key point is to find out the relation between a hyperspectral curve and a producing area environment, the property and characteristic components of the medicinal material.
Drawings
FIG. 1 is a flow chart of the present invention for predicting the content of para-hydroxybenzoic acid in Lycium barbarum fruit based on hyperspectral imaging spectrometer.
FIG. 2 shows an integrated device used in the present invention, in which the 1-metal frame, the 2-400-1000nm lens, the 3-1000-2400nm lens, the 4-halogen lamp, the 5-mobile platform, and the 6-instrument are provided with a computer.
Fig. 3 shows an original placement diagram of the medlar.
Fig. 4 is a region of interest extraction.
Detailed Description
The method of the present invention is illustrated by the following specific examples, but the present invention is not limited thereto, and any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included within the scope of the present invention.
The experimental procedures used in the following examples are all conventional procedures unless otherwise specified.
Materials, reagents and the like used in the following examples are commercially available unless otherwise specified.
The medlar adopted in the following embodiments is medlar of Ningqi No. 1, No. 5, No. 7 and No. 9 produced in Ningxia, Gansu, inner Mongolia, Xinjiang and Qinghai.
The method for chemically measuring the content of the medlar in the following examples is as follows:
and (3) determining the content of the p-hydroxybenzoic acid in the wolfberry fruit by using a high performance liquid chromatography triple quadrupole tandem mass spectrometry. Using an ACQUITY UPLC BEH C18 chromatographic column (100 mm. times.2.1 mm, 18 μm) at a column temperature of 40 deg.C; and (3) obtaining a corresponding extracted ion flow diagram by adopting an electrospray ion source and a negative ion detection mode, and quantifying by taking the peak area as an index. Extracting p-hydroxybenzoic acid from Lycium barbarum fruit under 70% ethanol and mobile phase conditions of 0.1% formic acid (A) -0.1% formic acid acetonitrile (B)
Examples 1,
The embodiment provides a model establishing method for predicting the content of p-hydroxybenzoic acid in wolfberry based on a hyperspectral imaging technology and a method for measuring the content of p-hydroxybenzoic acid in wolfberry, which comprises the following steps:
1) and 6 samples are taken, 75 samples are taken for each sample, and the samples are randomly divided into 5 groups and placed on a moving platform to be as far as possible not to exceed the range of a lens. When the wolfberry fruits are placed, the characteristics of each particle are highlighted, each group is arranged in a row, and a white board for black-white correction is placed at the position 5cm behind the sample. And waiting for instrument connection and self-checking. And setting scanning parameters of the hyperspectral imager, wherein the lens distance is 30cm, and the platform moving speed is 1.5 mm/s. The integration time of the 400-nm lens is set to be 4350 mus, and the frame time is 22000. The integration time for the 1000-plus 2400nm lens is 4000 mus, and the frame time is 35000. The placement of fructus Lycii is shown in figure 3. Performing spectrum scanning on the sample by using a hyperspectral imaging system, and collecting 400-1000nm and 1000-2400nm hyperspectral data to obtain a sample set spectrum;
2) the hyperspectral raw data is corrected by RAD Correction software carried by a spectrometer, and then the data is processed into relative reflectivity data by adopting a Flat Field Correction (Flat Field Correction) function in ENVI software.
3) And selecting the region of interest of the wolfberry fruit by utilizing ENVI. And exporting the average spectral value extracted from the region of interest and storing the average spectral value in a txt format. And (3) screening the initial relative reflectance value by a preprocessing method (preprocessing is carried out by three means of multivariate scattering correction, S-G smoothing and standard normalization, and the result judgment is carried out by using R, R square and adjusting R square, and the result is shown in table 1), and finally, the preprocessing method is determined to be multivariate scattering correction.
4) Correcting the relative reflectivity data by a multivariate scattering correction method; and introducing the corrected data into SPSS software, performing principal component analysis and dimensionality reduction, performing principal component regression analysis by using factors after dimensionality reduction, selecting the optimal factor number corresponding to the first time when the numerical value is stable and unchanged when a regression equation R and an R square are more than 0.99, and finally determining the optimal factor number to be 10 (the result is shown in a table 2).
5) And (3) performing dimensionality reduction on the obtained dimensionality reduction result according to the proportion of 7: and 3, dividing into a correction sample set and a test sample set.
6) Performing correlation analysis on the relative reflectivity data after the multivariate scattering correction and the content of the p-hydroxybenzoic acid (determined by a liquid-mass spectrometry method), and screening out wave bands with correlation numbers larger than 0.4 and extremely significant levels in the significance test to determine the wave bands as effective wave bands; namely 32-92 wave bands (wavelength range 556.686-990.096 nm).
7) Based on the screened most appropriate number of factors (10), respectively carrying out principal component analysis on the effective wave bands of the correction set and the verification sample set to remove redundant spectral information, and respectively obtaining main spectral information of the correction set and the verification sample set;
8) modeling the main spectral information of the correction set and the content of the p-hydroxybenzoic acid (measured by a liquid-mass combination method) by using a principal component regression method to obtain a prediction model of the content of the p-hydroxybenzoic acid in the medlar. The results of the model are shown in Table 3.
9) And (3) verification of the model: the main spectrum information of the verification sample set obtained in the step 6) is obtainedInputting a prediction model of the p-hydroxybenzoic acid content in the medlar, and calculating to obtain a calculated value of the p-hydroxybenzoic acid content in the verification sample set. Calculating R of the calculated value and the chemically measured value2Value and adjustment of R2The reliability of the model for predicting the p-hydroxybenzoic acid content in the obtained wolfberry was evaluated, and the results are shown in table 4. As can be seen from table 4, the error between the predicted value and the actual value is small, and the model accuracy is high.
Table 1 pretreatment method screening
TABLE 2 component factor number determination
TABLE 3 model coefficients
TABLE 4 coefficient of discrimination
Claims (3)
1. A model building method for predicting the content of p-hydroxybenzoic acid in wolfberry based on a hyperspectral imaging technology comprises the following steps:
1) establishing a sample spectrum:
collecting dried fructus Lycii products of different varieties and production places as sample set; performing spectrum scanning on the samples in the sample set by using a hyperspectral imaging system, and collecting 400-1000nm and 1000-2400nm hyperspectral data to obtain sample set spectra;
2) sample spectrum pretreatment:
a1) performing radiation calibration on original hyperspectral data of a sample;
b1) converting the data after radiation correction into relative reflectivity data by adopting a flat field processing method;
c1) processing the relative reflectivity data by a multivariate scattering correction method;
d1) performing principal component analysis and dimensionality reduction on the relative reflectivity data subjected to the multivariate scattering correction, performing principal component regression analysis by using factors subjected to dimensionality reduction, and selecting a factor number corresponding to the first occurrence of numerical stability and invariability when a regression equation R and an R square are more than 0.99 as an optimal factor number;
e1) carrying out correlation analysis on the relative reflectivity data after the multivariate scattering correction and the content of the p-hydroxybenzoic acid in the sample set measured by a chemical method, and screening out wave bands with correlation numbers larger than 0.4 and extremely significant levels in the significance test to determine the wave bands as effective wave bands;
f1) based on the optimal factor number screened out by d1), carrying out principal component analysis on the effective wave bands of the correction set and the verification sample set to remove redundant spectral information, and obtaining main spectral information;
3) establishing a correction model: modeling the main spectral information obtained by the calibration set spectrum and the content of the p-hydroxybenzoic acid contained in the sample determined by a chemical method by utilizing principal component regression to obtain a prediction model of the content of the p-hydroxybenzoic acid in the medlar;
the prediction model is shown in the following formula 1:
Y＝Z+AX1+BX2+CX3+DX4+EX5+FX6+GX7+HX8+IX9+JX10；
wherein, Z is 0.330, a is 0.059, B is 0.076, C is 0.043, D is 0.057, E is-0.050, F is-0.031, G is 0.010, H is-0.004, I is 0.006, J is 0.004; x1, X2 … … X10 are the 10 factors determined in step d 1);
in the step 1), the number of samples in the sample set is greater than or equal to 100; the hyperspectral imaging system is a HySpex series hyperspectral imaging spectrometer; the conditions for the spectral scan are as follows: the distance between a lens of the hyperspectral imaging spectrometer and the wolfberry fruit is 20-30cm, and the moving speed of a platform is 1.5 mm/s; the integration time was 4350 μ s when the collected spectral range was 400-1000nm, and the frame time was 18000; the integration time was 4500 μ s and the frame time was 46928 when the collected spectral range was 1000-2400 nm;
in the step 2), the number of the optimal factors is finally determined to be 10; in the step 2), the effective waveband is 28-108 waveband, and the corresponding wavelength range is 553 +/-3 nm-987 +/-3 nm;
in the step 2), SPSS software is adopted to carry out the principal component analysis.
2. The method of claim 1, wherein: in the step 3), the chemical method for measuring the content of the p-hydroxybenzoic acid in the sample is a liquid-mass combination method;
in the step 3), matlab software is adopted to establish the principal component regression model.
3. The method for predicting the content of parahydroxybenzoic acid in wolfberry comprising the modeling method for predicting the content of parahydroxybenzoic acid in wolfberry based on hyperspectral imaging technology as claimed in claim 1, comprising the steps of:
A) establishing a spectrum of a sample to be detected:
performing spectrum scanning on a sample to be detected by using a hyperspectral imaging system, and collecting 400-1000nm and 1000-2400nm hyperspectral data to obtain a sample set spectrum to be detected;
B) preprocessing the spectrum of a sample to be detected:
a2) performing radiation calibration on original hyperspectral data of a sample;
b2) converting the data after radiation calibration into relative reflectivity data by adopting a flat field processing method;
c2) processing the relative reflectivity data by a multivariate scattering correction method;
d2) screening the spectral information of the required effective waveband from the data processed in the step c2) according to the effective waveband determined in the step e 1); performing principal component analysis on the effective waveband to remove redundant spectral information to obtain main spectral information of the sample to be detected;
C) inputting the main spectral information of the sample to be detected into the formula 1, and calculating to obtain the content of p-hydroxybenzoic acid in the sample to be detected;
in the step A), the hyperspectral imaging system is a HySpex series hyperspectral imaging spectrometer;
the conditions for the spectral scan are as follows: the distance between a lens of the hyperspectral imaging spectrometer and the wolfberry fruit is 20-30cm, and the moving speed of a platform is 1.5 mm/s; the integration time was 4350 μ s when the collected spectral range was 400-1000nm, and the frame time was 18000; the integration time was 4500 μ s and the frame time was 46928 when the collected spectral range was 1000-2400 nm;
in the step B), the effective waveband is 28-108 waveband, and the corresponding wavelength range is 553 +/-3 nm-987 +/-3 nm.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811285081.9A CN109406419B (en) | 2018-10-31 | 2018-10-31 | Method for predicting content of p-hydroxybenzoic acid in wolfberry based on hyperspectral imaging technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811285081.9A CN109406419B (en) | 2018-10-31 | 2018-10-31 | Method for predicting content of p-hydroxybenzoic acid in wolfberry based on hyperspectral imaging technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109406419A CN109406419A (en) | 2019-03-01 |
| CN109406419B true CN109406419B (en) | 2020-09-11 |
Family
ID=65470820
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811285081.9A Active CN109406419B (en) | 2018-10-31 | 2018-10-31 | Method for predicting content of p-hydroxybenzoic acid in wolfberry based on hyperspectral imaging technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109406419B (en) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103411973B (en) * | 2013-09-03 | 2016-02-24 | 西北农林科技大学 | A kind of based on anthocyanin content method for measuring in the vinifera pericarp of EO-1 hyperion |
| CN103558167B (en) * | 2013-10-31 | 2016-08-17 | 华南理工大学 | A kind of method of sodium chloride content in quick mensuration butcher's meat |
| CN105241822A (en) * | 2015-08-28 | 2016-01-13 | 河南科技大学 | Measurement method of content of anthocyanin in leaves of peony on the basis of hyperspectrum |
-
2018
- 2018-10-31 CN CN201811285081.9A patent/CN109406419B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN109406419A (en) | 2019-03-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Weng et al. | Hyperspectral imaging for accurate determination of rice variety using a deep learning network with multi-feature fusion | |
| Zhao et al. | Near infrared reflectance spectroscopy for determination of the geographical origin of wheat | |
| Li et al. | Authenticity identification and classification of Rhodiola species in traditional Tibetan medicine based on Fourier transform near-infrared spectroscopy and chemometrics analysis | |
| Huang et al. | Improved generalization of spectral models associated with Vis-NIR spectroscopy for determining the moisture content of different tea leaves | |
| CN109540831B (en) | Wolfberry variety identification method based on hyperspectral imaging technology | |
| CN108169165B (en) | Maltose mixture quantitative analysis method based on terahertz spectrum and image information fusion | |
| CN113008817A (en) | Method for rapidly identifying authenticity and quality of bitter apricot kernels based on hyperspectral imaging technology | |
| CN113762208B (en) | Spectrum conversion method of near infrared spectrum and characteristic spectrum and application thereof | |
| CN104020128A (en) | Method for rapidly identifying propolis source | |
| Cui et al. | Identification of maize seed varieties based on near infrared reflectance spectroscopy and chemometrics | |
| CN107132194A (en) | A kind of pseudo-ginseng and its adulterant discrimination method based on uv-vis spectra and Chemical Pattern Recognition | |
| Wang et al. | Extraction and classification of origin characteristic peaks from rice Raman spectra by principal component analysis | |
| Sun et al. | Water content detection of potato leaves based on hyperspectral image | |
| Wang et al. | Intelligent detection of hard seeds of snap bean based on hyperspectral imaging | |
| CN113655027A (en) | Method for rapidly detecting tannin content in plant by near infrared | |
| CN114112983A (en) | Python data fusion-based Tibetan medicine all-leaf artemisia rupestris L producing area distinguishing method | |
| CN109406419B (en) | Method for predicting content of p-hydroxybenzoic acid in wolfberry based on hyperspectral imaging technology | |
| CN109406421B (en) | Method for predicting ferulic acid content in wolfberry fruit based on hyperspectral imaging technology | |
| CN109406420B (en) | Method for predicting content of scopoletin in fructus lycii based on hyperspectral imaging technology | |
| Sun et al. | Nondestructive identification of barley seeds varieties using hyperspectral data from two sides of barley seeds | |
| Chen et al. | Monitoring green tea fixation quality by intelligent sensors: comparison of image and spectral information | |
| Xu et al. | Hyperspectral imaging with machine learning for non-destructive classification of Astragalus membranaceus var. mongholicus, Astragalus membranaceus, and similar seeds | |
| CN109540837A (en) | The method that near-infrared quickly detects Boehmeria nivea leaves wood fibre cellulose content | |
| Li et al. | The prediction model of nitrogen nutrition in cotton canopy leaves based on hyperspectral visible‐near infrared band feature fusion | |
| CN117132843B (en) | Wild ginseng, under-forest mountain ginseng and garden ginseng in-situ identification method, system and related equipment |
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 |