Traditional Chinese medicine organic acid content prediction method and system based on hyperspectral imaging

Abstract

The invention provides a hyperspectral imaging-based traditional Chinese medicine organic acid content prediction method and system, and relates to the technical field of traditional Chinese medicine analysis and detection. The method comprises the following steps: acquiring the organic acid content of a traditional Chinese medicine sample, acquiring specific wave band reflectivity by using a hyperspectral imaging technology to construct a traditional Chinese medicine data set, calculating wavelengths, calculating wavelength derivatives, screening wavelengths possibly associated with the organic acid content, and grouping to establish an associated data set; and constructing a polynomial regression model by using the screened wavelengths possibly associated with the organic acid and the corresponding reflectance, and performing optimization through a correlation coefficient formula so as to predict the content of the organic acid, namely counting the proportion of the number of samples of which the residual error exceeds a threshold value for traditional Chinese medicines of which the actual content of the organic acid is different from the predicted value of the model, and determining the content of the organic acid in the traditional Chinese medicines. A model is independently constructed for traditional Chinese medicine types exceeding a proportion threshold value to predict the organic acid content, a prediction system is perfected, and a lossless and accurate solution is provided for detection of the organic acid content of traditional Chinese medicine.

Description

Technical Field

[0001] The present invention relates to the technical field of traditional Chinese medicine analysis and detection, and in particular to a method and system for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging. Background Art

[0002] As the key active ingredients or indicator ingredients in traditional Chinese medicine, the precise determination of organic acids plays a vital role in the quality control of traditional Chinese medicine, the analysis of pharmacological mechanisms, and the process of innovative research and development of new drugs. Roots, rhizomes, and stems of traditional Chinese medicine are widely used in many traditional prescriptions and are used in large quantities. The level of their organic acid content is closely related to the efficacy and quality of traditional Chinese medicine. For example, some organic acids in roots and rhizomes of traditional Chinese medicine may participate in regulating the body's immune function and antioxidant processes; organic acids in stems of traditional Chinese medicine may have important contributions to their cardiovascular activity, antibacterial and anti-inflammatory properties. However, traditional methods for determining the content of organic acids in traditional Chinese medicine, such as chemical titration and high-performance liquid chromatography, often have defects such as cumbersome operation, long time consumption, the need for a large amount of chemical reagents, and certain destructiveness to samples. It is difficult to meet the needs of large-scale, rapid and non-destructive testing, resulting in restrictions on the comprehensive and efficient evaluation of the content of organic acids in traditional Chinese medicine. In addition, although some emerging detection technologies have been improved to a certain extent, they still have problems such as low detection sensitivity, low specificity, expensive equipment, and complex maintenance, which limit their widespread promotion and application in the traditional Chinese medicine industry. Therefore, developing a new, accurate, efficient, and non-destructive method for predicting the content of organic acids in traditional Chinese medicine has become a technical problem that needs to be overcome in the field of modern research on traditional Chinese medicine.

[0003] As an advanced spectral analysis method, hyperspectral imaging technology has been widely used in agriculture, food, environment and other fields in recent years and has shown great potential. It can simultaneously obtain the reflectance information of the sample in the continuous spectral range, thereby reflecting the complex chemical composition and physical structure characteristics inside the sample. In the field of traditional Chinese medicine research, its application is still in the initial and exploratory stage. The present invention constructs a method and system for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging technology, making full use of the spectral information collection capability of hyperspectral imaging, focusing on roots, rhizomes, and stems of traditional Chinese medicine, and accurately determining the organic acid content of each traditional Chinese medicine sample, thereby establishing a representative and comprehensive organic acid content data base, and realizing rapid, non-destructive, and accurate prediction of the organic acid content of traditional Chinese medicine.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, the specific steps comprising:

[0008] Step 1: Obtain the content of organic acids in each TCM sample to be measured, use hyperspectral imaging technology, focus on the 400 to 2500 nm band, obtain the reflectivity of different wavelengths in each TCM sample, and construct a TCM data set with different wavelengths and corresponding reflectivity in each TCM;

[0009] Step 2: Calculate the first-order and second-order derivatives of each wavelength, screen the wavelengths that may be associated with organic acids based on the derivative results, establish an organic acid-related data set using the wavelengths that may be associated with organic acids and the corresponding reflectance, and group the data according to the name of the traditional Chinese medicine;

[0010] Step 3: Extract data from the organic acid-related data set, use the wavelengths that may be associated with organic acids and the corresponding reflectance to build a polynomial regression model, and select the polynomial regression model with the largest correlation value through the correlation coefficient formula to predict the organic acid content of traditional Chinese medicine;

[0011] Step 4: Calculate the residuals of each Chinese medicine sample. For each Chinese medicine individual, count the number of Chinese medicine samples whose residuals exceed the preset residual threshold, and then calculate the proportion of this number in the total number of Chinese medicine samples of the same Chinese medicine. When the proportion exceeds the preset proportion threshold, mark this Chinese medicine as a special Chinese medicine. The residual is the difference between the organic acid content actually observed in the Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model.

[0012] Step 5: A polynomial regression model is constructed separately for the organic acid content of the special Chinese medicine and the reflectance of the corresponding wavelength to predict the organic acid content of the special Chinese medicine.

[0013] Furthermore, the Chinese medicine sample was ground into fine powder, distilled water was added to fully soak the powder and then stirred for 30 minutes, solid impurities in the extract were removed by filter paper, distilled water was added, and titration was performed with 0.1 mol / L sodium hydroxide solution. When the color of the solution changed from colorless to pink and did not fade within half a minute, the volume of the sodium hydroxide solution consumed was recorded, and the amount of sodium hydroxide consumed was calculated by the formula, that is, the amount of organic acid, and the content of organic acid in the Chinese medicine sample was calculated.

[0014] Further, the amount of sodium hydroxide consumed is calculated using the formula n=c·V, where n represents the amount of sodium hydroxide, c represents the molar concentration of the sodium hydroxide solution, i.e., 0.1 mol / L, and V represents the volume of the sodium hydroxide solution added dropwise to the extract. The content of organic acids in the traditional Chinese medicine samples was calculated, where w represented the content of organic acids in the traditional Chinese medicine samples, M represented the molar mass of organic acids, and m represented the mass of traditional Chinese medicine samples.

[0015] Furthermore, hyperspectral imaging technology was used, the spectral range was pre-set to 400-2500nm, the resolution was set to 10nm, and a hyperspectral image data cube was constructed. The reflected light signal intensity of each pixel on the surface of the traditional Chinese medicine at each wavelength was directly obtained through the spectroscopic component and the detector, and the average reflected light signal intensity of all pixel points was calculated to construct the reflectance formula:

[0016]

[0017] In the formula, R is the reflectivity, I yb It is expressed as the reflected light signal intensity of a standard white plate under the same lighting conditions and wavelength, I bb It is expressed as the reflected light signal intensity of the Chinese medicine sample at a certain wavelength.

[0018] Furthermore, for each Chinese medicine sample, the reflectance data at each wavelength is obtained according to the pre-set spectral range of 400nm to 2500nm and the resolution of 10nm, and all the wavelength data corresponding to each Chinese medicine sample are taken as a row. The values ​​of each column in the row are the reflectance of the Chinese medicine sample at the corresponding wavelength. Different rows of Chinese medicine sample data are arranged in sequence to construct a two-dimensional Chinese medicine data set, in which rows represent different Chinese medicine sample individuals, and columns represent different wavelengths and their corresponding reflectance data.

[0019] Further, extract data from the data collection table, and pre-set the reflectivity sequence as R, R = {r 400 , r 410 ,…,r 2400 , r 2500}, where r refers to the reflectivity at each wavelength, and the wavelength sequence is set to λ, λ={λ 400 ,λ 410 , …, λ 2400 ,λ 2500}, indicating the wavelength value starting from 400nm and extending to 2500nm with an interval of 10nm, and setting the wavelength interval to Δλ. Since the resolution is 10nm, Δλ=10;

[0020] Using the central difference method, calculate the first derivative sequence of the reflectivity sequence R with respect to the wavelength λ:

[0021]

[0022] Where i represents the wavelength, i = 400, 410, ..., 2500, R ′ (λ i ) represents the first-order derivative sequence of the reflectivity sequence R with respect to the wavelength λ, that is, at the wavelength λ i The rate of change of reflectivity at wavelength λ = 400 and 2500 is calculated by the forward difference method:

[0023]

[0024] In the formula, R ′ (λ 400 ) is expressed as the first-order derivative of the reflectivity series R with respect to the wavelength of 400nm, R ′ (λ 400 ) is expressed as the first-order derivative of the reflectivity series R with respect to the wavelength of 2500nm,

[0025] Based on the first-order derivative sequence, the second-order derivative is calculated using the central difference method:

[0026]

[0027] In the formula, R″(λ i ) represents the reflectivity sequence R with respect to wavelength λ at wavelength value λ i The value of the second-order derivative sequence at ,

[0028] The calculation of the second-order derivative of the boundary point is similar to the processing of the first-order derivative boundary point, requiring a special one-sided difference formula:

[0029]

[0030] In the formula, R″(λ 400 ) represents the reflectivity sequence R with respect to the wavelength value λ 400 The value of the second-order derivative sequence at , R″(λ 2500 ) represents the reflectivity sequence R with respect to the wavelength value λ 2500 The value of the second-order derivative sequence at ,

[0031] The position where the first-order derivative is zero and the second-order derivative is less than zero is defined as the absorption peak position, and the wavelength corresponding to the absorption peak position is recorded. At a certain wavelength, the number of Chinese medicine samples corresponding to its absorption peak reaches 90% or more of the total number of Chinese medicine samples. This wavelength can be preliminarily identified as a wavelength that may be associated with organic acids.

[0032] Furthermore, we first extracted the data of the same Chinese medicine in the organic acid-related data set, and used the reflectance corresponding to the wavelengths that were initially identified as possibly associated with organic acids as independent variables, and the corresponding organic acid content data of the same Chinese medicine as the dependent variable to construct a quadratic polynomial regression model:

[0033]

[0034] Where x represents the wavelength reflectance data, is expressed as the predicted organic acid content data, β0, β1 and β2 are expressed as coefficients,

[0035] The parameters β0, β1 and β2 in the model are determined by the least squares method, that is, the parameters β0, β1 and β2 in the model are determined by minimizing the sum of squared errors:

[0036]

[0037] In the formula, SSE means minimizing the sum of squared errors, n is the number of TCM samples, x a is the reflectance of the ath TCM sample at the wavelength that is preliminarily identified as possibly associated with organic acids, y a is the organic acid content of the a-th Chinese medicine sample, a=1, 2, …, z, where z is the number of Chinese medicines collected, and z is a positive integer ≥ 50. The partial derivatives with respect to β0, β1 and β2 are calculated respectively, and the partial derivatives are set equal to zero to solve the quadratic polynomial regression model;

[0038] For β0:

[0039]

[0040] In the formula, It means to find the partial derivative of β0;

[0041] For β1:

[0042]

[0043] In the formula, It means to find the partial derivative of β1;

[0044] For β2:

[0045]

[0046] In the formula, It means to find the partial derivative of β2;

[0047] By solving the above system of three partial derivative equations, we can obtain the values ​​of β0, β1 and β2;

[0048] Then, the correlation coefficient between the reflectance data and the organic acid content data under the model was calculated;

[0049] The correlation coefficient calculation formula is as follows:

[0050]

[0051] In the formula, r represents the correlation coefficient between the reflectance data and the organic acid content data, is the mean value of the reflectance data corresponding to the wavelength that is preliminarily identified as possibly associated with organic acids. is the mean of the organic acid content data,

[0052] Repeat the above process of building the model and calculating the correlation coefficient to obtain multiple different quadratic polynomial regression models and their corresponding correlation coefficients;

[0053] Finally, these correlation coefficients were compared, and the wavelength with the highest correlation coefficient value r was selected and determined as the specific wavelength related to the content of organic acids in traditional Chinese medicine. The polynomial regression model with the largest correlation value was selected to predict the content of organic acids in traditional Chinese medicine.

[0054] Furthermore, for each Chinese medicine sample, the residual is calculated separately, and the residual calculation formula is:

[0055]

[0056] In the formula, represents the organic acid content calculated by the quadratic polynomial regression model, e a It represents the difference between the organic acid content actually observed in each Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model.

[0057] The residual threshold is pre-set to e max , the ratio threshold is q max , for each Chinese medicine, the residual e is counted a Greater than or equal to the residual threshold e max The number of TCM samples is calculated, and then the proportion of the number of TCM samples that meet the above conditions in the total number of TCM samples of this type of TCM is calculated. If it exceeds the proportion threshold q max , marking this Chinese medicine as a special Chinese medicine.

[0058] Furthermore, for the types marked as special Chinese medicines, the extracted data include: Chinese medicine name, actual organic acid content, specific wavelength reflectance,

[0059] The specific wavelength reflectance data x is used as the independent variable, and the corresponding organic acid content data y is used as the dependent variable to construct a quadratic polynomial regression model:

[0060]

[0061] The parameters β0, β1 and β2 in the model are determined by the least squares method, that is, the parameters β0, β1 and β2 in the model are determined by minimizing the sum of squared errors:

[0062]

[0063] Calculate the partial derivatives with respect to β0, β1 and β2 respectively, and set the partial derivatives equal to zero to solve the quadratic polynomial regression model.

[0064] The present invention further provides a system for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, wherein the system is used to execute the above-mentioned method for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, comprising:

[0065] The data set acquisition module is used to obtain the content of organic acids in each Chinese medicine sample to be measured, and use hyperspectral imaging technology to focus on the 400 to 2500nm band to obtain the reflectivity of different wavelengths in each Chinese medicine sample, and construct a Chinese medicine data set with different wavelengths and corresponding reflectivity in each Chinese medicine;

[0066] The data grouping module is used to calculate the first-order and second-order derivatives of each wavelength, screen the wavelengths that may be associated with organic acids based on the derivative results, establish an organic acid-related data set using the wavelengths that may be associated with organic acids and the corresponding reflectance, and group the data according to the name of the traditional Chinese medicine;

[0067] The linear regression module is used to extract data from the organic acid-related data set, build a polynomial regression model using the wavelengths that may be associated with organic acids and the corresponding reflectance, and select the polynomial regression model with the largest correlation value through the correlation coefficient formula to predict the organic acid content of traditional Chinese medicine;

[0068] The difference analysis module is used to calculate the residuals of each Chinese medicine sample. For each Chinese medicine individual, the number of Chinese medicine samples whose residuals exceed the preset residual threshold is counted, and then the proportion of this number in the total number of Chinese medicine samples of the same Chinese medicine is calculated. When the proportion exceeds the preset proportion threshold, the Chinese medicine is marked as a special Chinese medicine. The residual is the difference between the organic acid content actually observed in the Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model;

[0069] The special content analysis module is used to construct a polynomial regression model based on the organic acid content of special Chinese medicine and the reflectivity of the corresponding wavelength, so as to predict the organic acid content of special Chinese medicine.

[0070] Compared with the prior art, the present invention has the following beneficial effects:

[0071] The present invention adopts hyperspectral imaging technology, focuses on the 400 to 2500nm band, can quickly obtain the reflectance data of traditional Chinese medicine samples, construct a traditional Chinese medicine data set, calculate the derivative of each wavelength to accurately locate the wavelength that may be associated with organic acids, establish an organic acid-related data set and group it by the name of traditional Chinese medicine, and conduct targeted analysis and processing for subsequent prediction of organic acid content of traditional Chinese medicine, reducing the interference of irrelevant data. Data is extracted from the organic acid-related data set to construct a polynomial regression model, and the optimal model is screened out according to the correlation coefficient formula, so as to achieve accurate prediction of organic acid content in traditional Chinese medicine, and avoid the problem of composition change and structural damage that may be caused by destroying traditional Chinese medicine samples.

[0072] Through residual calculation and proportion statistics, we can accurately identify those special Chinese medicine individuals that show special patterns in the relationship between spectrum and organic acid content and have poor adaptability to the overall model. Polynomial regression models are constructed for these special Chinese medicines separately, which can maintain good prediction performance when facing Chinese medicines with different characteristics. BRIEF DESCRIPTION OF THE DRAWINGS

[0073] Figure 1 It is a schematic diagram of the overall method flow of the present invention;

[0074] Figure 2 It is a schematic diagram of the overall system module of the present invention. DETAILED DESCRIPTION

[0075] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0076] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The words "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0077] Example:

[0078] See also Figure 1 , the present invention provides a technical solution:

[0079] A method for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, the specific steps comprising:

[0080] Step 1: Obtain the content of organic acids in each TCM sample to be measured, use hyperspectral imaging technology, focus on the 400 to 2500 nm band, obtain the reflectivity of different wavelengths in each TCM sample, and construct a TCM data set with different wavelengths and corresponding reflectivity in each TCM;

[0081] Grind each Chinese medicine sample into fine powder, take an appropriate amount of powder and place it in a container, add distilled water to soak the powder, and then stir for 30 minutes. Next, filter the extract to remove solid impurities, add distilled water to dilute, and use 0.1 mol / L sodium hydroxide solution for titration. Slowly add sodium hydroxide solution while stirring continuously until the color of the solution changes from colorless to pink and does not fade within half a minute. Record the volume of sodium hydroxide solution consumed at this time.

[0082] For example, for a Chinese angelica sample, the volume of sodium hydroxide solution consumed in the titration process is 30.00 mL. According to the formula n = c·V, the amount of sodium hydroxide consumed is calculated to be 0.003 mol, which is the amount of organic acid. Assuming that the mass of the Chinese angelica sample is 6.00 g, and the molar mass of organic acid in angelica is known to be M = 120 g / mol, according to the formula The content of organic acid in the Chinese angelica root sample was calculated to be 0.6%.

[0083] Establish accurate basic data for determination of organic acid content to provide reliable dependent variable data for subsequent correlation analysis and model building, thereby analyzing the relationship between it and spectral reflectance.

[0084] The 400-2500nm band range can cover the absorption and reflection characteristics of light by various chemical components in traditional Chinese medicine. Among them, the carboxyl and hydroxyl groups of organic acids have corresponding absorption characteristics in this band, and spectral absorption and reflection behaviors will occur. By acquiring reflectance data, these wavelengths and reflectance information related to the organic acid content can be captured. By establishing a quantitative relationship model between the spectrum and the organic acid content, a large number of samples can be quickly detected and analyzed in a short time without the need for complex pretreatment or destruction of new samples.

[0085] The PFD VNI R hyperspectral imaging system was used, with the spectral range pre-set to 400-2500nm and the resolution set to 10nm. The processed Chinese medicine samples were placed one by one in a fixed position of the imaging device to construct a hyperspectral image data cube. The reflected light signal intensity of each pixel point on the surface of the Chinese medicine at each wavelength was directly obtained through the spectroscopic component and the detector. First, the average value of the reflected light signal intensity of all pixel points of the Chinese medicine sample at a wavelength of 400nm was calculated, and then the reflectance formula was constructed according to: (where R represents reflectivity, I bb It is expressed as the reflected light signal intensity of the standard white plate under the same lighting conditions and wavelength, I yb The reflectance at 400nm is calculated (expressed as the reflected light signal intensity of the Chinese medicine sample at different wavelengths). Then, the reflectance at 410nm, 420nm to 2500nm is calculated in sequence according to the same method and steps, so as to obtain the reflectance data of the Chinese medicine sample in the entire set spectral range.

[0086] Afterwards, the reflectance data of each TCM sample at each wavelength from 400nm to 2500nm is taken as a row, and the value of each column in the row is the reflectance of the TCM sample at the corresponding wavelength. The TCM sample data rows are arranged in sequence according to the type of TCM to construct a two-dimensional TCM data set, in which rows represent different TCM sample individuals, and columns represent different wavelengths and their corresponding reflectance data. The next step will be to perform wavelength correlation analysis based on these data.

[0087] Step 2: Calculate the first-order and second-order derivatives of each wavelength, screen the wavelengths that may be associated with organic acids based on the derivative results, establish an organic acid-related data set using the wavelengths that may be associated with organic acids and the corresponding reflectance, and group the data according to the name of the traditional Chinese medicine;

[0088] Extract data from the data collection table and pre-set the reflectivity sequence to R, R = {r 400 , r 410 ,…,r 2400 , r 2500}, where r refers to the reflectivity at different wavelengths, for example r 400 , r 410 is the reflectivity at 400nm and 410nm wavelengths, and the wavelength sequence is set to λ, λ={λ 400 ,λ 410 , …, λ 2400 ,λ 2500}, indicating the wavelength value starting from 400nm and extending to 2500nm with an interval of 10nm, and the wavelength interval is set to Δλ. Since the resolution is 10nm, Δλ=10 here.

[0089] Using the central difference method, calculate the first derivative sequence of the reflectivity sequence R with respect to the wavelength λ:

[0090]

[0091] Where i represents the wavelength, i = 400, 410, ..., 2500, R ′ (λ i ) represents the first-order derivative sequence of the reflectivity sequence R with respect to the wavelength λ, that is, at the wavelength λ i The rate of change of reflectivity at wavelength λ = 400 and 2500 is calculated by the forward difference method:

[0092]

[0093] In the formula, R ′ (λ 400 ) is expressed as the first-order derivative of the reflectivity series R with respect to the wavelength of 400nm, R ′ (λ 400 ) is expressed as the first-order derivative of the reflectivity series R with respect to the wavelength of 2500nm.

[0094] Calculating the first-order derivative sequence of the reflectivity sequence R with respect to the wavelength λ can reflect the changing trend of the reflectivity with wavelength and help determine the turning points of the reflectivity change, which may be related to important spectral features such as absorption peaks.

[0095] Based on the first-order derivative sequence, the second-order derivative is calculated using the central difference method:

[0096]

[0097] In the formula, R″(λ i ) represents the reflectivity sequence R with respect to the wavelength value λ i The value of the second derivative sequence at .

[0098] When the first-order derivative is zero and the second-order derivative is less than zero, it indicates that the reflectivity at this wavelength changes from rising to falling, forming a local maximum point, which is consistent with the characteristics of the absorption peak.

[0099] The calculation of the second-order derivative of the boundary point is similar to the processing of the first-order derivative boundary point, requiring a special one-sided difference formula:

[0100]

[0101] In the formula, R″(λ 400 ) represents the reflectivity sequence R with respect to the wavelength value λ 400 The value of the second-order derivative sequence at , R″(λ 2500 ) represents the reflectivity sequence R with respect to the wavelength value λ2500 The value of the second derivative sequence at .

[0102] The absorption peak is located according to the results of the first-order derivative and the second-order derivative. The position where the first-order derivative is zero and the second-order derivative is less than zero is defined as the absorption peak position:

[0103] Satisfy R ′ (λ i )=0 and R″(λ i )<0 wavelength λ i Defined as the absorption peak position.

[0104] The absorption peak is located according to the results of the first-order derivative and the second-order derivative. This is based on the principle of spectroscopy. When light is irradiated on a substance, the chemical bonds or functional groups in the substance will absorb light of a certain wavelength. Organic acids contain specific functional groups, and these functional groups will have characteristic absorption peaks within a certain wavelength range. This causes the reflectivity to decrease at this wavelength, which is manifested as an absorption peak on the spectrum curve. The conditions that the first-order derivative is zero and the second-order derivative is less than zero can accurately identify the peak point of this reflectivity decrease.

[0105] Construct an organic acid association data set. At a certain wavelength, the number of Chinese medicine samples with absorption peaks reaches 90% or more of the total number of Chinese medicine samples, that is, the wavelength preliminarily identified as possibly associated with organic acids. The reflectivity corresponding to the wavelength preliminarily identified as possibly associated with organic acids is extracted and filled into the organic acid association data set.

[0106] If 90% of the TCM samples have an absorption peak at a certain wavelength, then this wavelength is likely to be a universal characteristic wavelength related to organic acids. This helps to screen out wavelengths that are truly closely related to organic acids, rather than false associations caused by the special ingredients of individual TCM samples or accidental factors.

[0107] The rows of the organic acid association data set represent different Chinese medicine samples, and the columns contain the name of the Chinese medicine, the organic acid content, and the reflectance corresponding to each wavelength associated with the organic acid. The data are grouped according to the name of the Chinese medicine, and within each Chinese medicine group, the Chinese medicine samples are sorted in ascending order according to the organic acid content.

[0108] For the same Chinese medicine samples, their organic acid content data are arranged in order from low to high. After completing the operation of grouping the data by the name of the Chinese medicine, the Chinese medicine samples are sorted in ascending order according to the numerical value of the organic acid content within each Chinese medicine group. Through the ascending sorting method, it is possible to more intuitively observe the corresponding change law of the reflectivity data when the organic acid content of the same Chinese medicine changes. In the subsequent processing process, for the types of Chinese medicines that are calibrated as special Chinese medicines, it is more convenient to build a separate model, thereby analyzing the relationship between organic acid content and reflectivity, and improving the method of predicting organic acid content in Chinese medicines.

[0109] Step 3: Extract data from the organic acid-related data set, use the wavelengths that may be associated with organic acids and the corresponding reflectance to build a polynomial regression model, and select the polynomial regression model with the largest correlation value through the correlation coefficient formula to predict the organic acid content of traditional Chinese medicine;

[0110] Extract the data from the organic acid-related data set, use the reflectance data corresponding to the wavelengths that are initially identified as possibly associated with organic acids as independent variables, and the corresponding organic acid content data as the dependent variable, to construct a quadratic polynomial regression model:

[0111] The wavelength reflectance data x is used as the independent variable and the corresponding organic acid content data y is used as the dependent variable to construct a quadratic polynomial regression model:

[0112]

[0113] The quadratic polynomial model was chosen because there is a nonlinear relationship between reflectance and organic acid content, and the quadratic term can capture this curvilinear relationship.

[0114] The parameters β0, β1 and β2 in the model are determined by the least squares method, that is, the parameters β0, β1 and β2 in the model are determined by minimizing the sum of squared errors:

[0115]

[0116] In the formula, SSE means minimizing the sum of squared errors, n is the number of TCM samples, and x a is the reflectance of the ath TCM sample at the wavelength that is preliminarily identified as possibly associated with organic acids, y ais the organic acid content of the a-th Chinese medicine sample, a=1, 2, ..., z, where z is the number of Chinese medicines collected, and z is a positive integer ≥ 1. The model obtained in this way can fit the data in the sense of minimizing the error, because in actual measurement, due to the influence of various factors, there must be errors between the measured value and the true value. The least squares method is to find an optimal combination of model parameters to minimize the sum of squares of these errors, so that the model can best fit the existing data and improve the accuracy of the prediction.

[0117] Calculate the partial derivatives with respect to β0, β1 and β2 respectively, and set the partial derivatives equal to zero to solve the quadratic polynomial regression model.

[0118] For β0:

[0119]

[0120] In the formula, It means to find the partial derivative of β0;

[0121] For β1:

[0122]

[0123] In the formula, It means to find the partial derivative of β1;

[0124] For β2:

[0125]

[0126] In the formula, It means to find the partial derivative of β2;

[0127] By solving the above system of three partial derivative equations, the values ​​of β0, β1 and β2 can be obtained.

[0128] Then, the correlation coefficient between the reflectance data and the organic acid content data under this model is calculated. The correlation coefficient calculation formula is as follows:

[0129]

[0130] In the formula, r represents the correlation coefficient between the reflectance data and the organic acid content data, is the mean value of the reflectance data corresponding to the wavelength that is preliminarily identified as possibly associated with organic acids. is the mean of the organic acid content data.

[0131] The above process of constructing the model and calculating the correlation coefficient is repeated to obtain multiple different quadratic polynomial regression models and their corresponding correlation coefficients.

[0132] Finally, the correlation coefficients were compared, and the wavelength with the highest correlation coefficient was selected as the specific wavelength related to the organic acid content of traditional Chinese medicine. The polynomial regression model with the largest correlation value was selected to predict the organic acid content of traditional Chinese medicine. This is done because the larger the correlation coefficient, the stronger the linear relationship between the independent variable (reflectivity data) and the dependent variable (organic acid content data) of the model, and the better the model's fit and prediction ability to the data, so that the organic acid content of traditional Chinese medicine can be predicted more accurately.

[0133] Step 4: Calculate the residuals of each Chinese medicine sample. For each Chinese medicine individual, count the number of Chinese medicine samples whose residuals exceed the preset residual threshold, and then calculate the proportion of this number in the total number of Chinese medicine samples of the same Chinese medicine. When the proportion exceeds the preset proportion threshold, mark this Chinese medicine as a special Chinese medicine. The residual is the difference between the organic acid content actually observed in the Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model.

[0134] After completing the construction of the quadratic polynomial regression model and determining the model parameters β0, β1, and β2 of traditional Chinese medicine, the residuals were calculated for each traditional Chinese medicine sample:

[0135]

[0136] In the formula, represents the organic acid content calculated by the quadratic polynomial regression model, e a It represents the difference between the organic acid content actually observed in each TCM sample and the organic acid content calculated by the quadratic polynomial regression model.

[0137] Preset the residual threshold e max =5% and the ratio threshold q max =95%. After a large amount of experimental data verification and analysis, it was found that when the residual threshold was set to 5% and the proportion threshold was set to 95%, special Chinese medicines with large differences from the overall model fitting could be effectively screened out, while ensuring the stability and accuracy of the model.

[0138] Each Chinese medicine sample is treated as a group. For each group of Chinese medicine, the residual e in each group of Chinese medicine is counted. a The number of TCM samples that is greater than or equal to 5% is calculated as the proportion of the total number of TCM samples in this group. If it exceeds the proportion threshold q max , marking this group of Chinese medicines as special types of Chinese medicines.

[0139] By calculating the residuals, the fitting effect of the quadratic polynomial regression model on each Chinese medicine sample can be evaluated. The smaller the residuals, the more accurate the model's prediction of the Chinese medicine sample; conversely, if the residuals are too large, it means that the model does not fit the Chinese medicine sample well. Therefore, by identifying the proportion of Chinese medicine samples with residuals exceeding the threshold, it is possible to find those Chinese medicine types that differ greatly from the overall model fitting trend, so as to analyze the organic acid content of special Chinese medicines separately in the future.

[0140] Step 5: A polynomial regression model is constructed based on the organic acid content of the special Chinese medicine and the reflectance of the corresponding wavelength to predict the organic acid content of the special Chinese medicine;

[0141] Due to their particularity, special Chinese medicines cannot be well described by models applicable to general Chinese medicines. Building a separate model can more accurately capture the intrinsic relationship between the spectra of special Chinese medicines and the organic acid content, improve the accuracy of predicting the organic acid content of special Chinese medicines, and enable both conventional and special Chinese medicines to have a suitable model for predicting the organic acid content.

[0142] For the types marked as special Chinese medicines, the extracted data include: name of the Chinese medicine, actual organic acid content, and reflectivity at specific wavelengths.

[0143] Extracting the names of Chinese medicines can accurately filter out the data of special Chinese medicines and avoid interference from other non-special Chinese medicine data. Once it is identified as a special medicine, the model pre-built for the special medicine can be directly called to analyze the organic acid content. For example, if a special Chinese medicine is ginseng, all reflectance and organic acid content data related to ginseng can be filtered out from the entire data set based on the name ginseng to build a quadratic polynomial regression model specifically for ginseng.

[0144] The specific wavelength reflectance data x is used as the independent variable, and the corresponding organic acid content data y is used as the dependent variable to construct a quadratic polynomial regression model:

[0145]

[0146] Although special Chinese medicines have different characteristics from general Chinese medicines, there is an overall nonlinear relationship between reflectance and organic acid content, and the quadratic polynomial regression model can capture this nonlinear relationship.

[0147] The parameters β0, β1 and β2 in the model are determined by the least squares method, that is, the parameters β0, β1 and β2 in the model are determined by minimizing the sum of squared errors:

[0148]

[0149] Calculate the partial derivatives with respect to β0, β1 and β2 respectively, and set the partial derivatives equal to zero to solve the quadratic polynomial regression model.

[0150] See also Figure 2 The present invention also provides a system for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, wherein the system is used to execute the above-mentioned method for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, comprising:

[0151] The data set acquisition module is used to obtain the content of organic acids in each Chinese medicine sample to be measured, and use hyperspectral imaging technology to focus on the 400 to 2500nm band to obtain the reflectivity of different wavelengths in each Chinese medicine sample, and construct a Chinese medicine data set with different wavelengths and corresponding reflectivity in each Chinese medicine;

[0152] The data grouping module is used to calculate the first-order and second-order derivatives of each wavelength, screen the wavelengths that may be associated with organic acids based on the derivative results, establish an organic acid-related data set using the wavelengths that may be associated with organic acids and the corresponding reflectance, and group the data according to the name of the traditional Chinese medicine;

[0153] The linear regression module is used to extract data from the organic acid-related data set, build a polynomial regression model using the wavelengths that may be associated with organic acids and the corresponding reflectance, and select the polynomial regression model with the largest correlation value through the correlation coefficient formula to predict the organic acid content of traditional Chinese medicine;

[0154] The difference analysis module is used to calculate the residuals of each Chinese medicine sample. For each Chinese medicine individual, the number of Chinese medicine samples whose residuals exceed the preset residual threshold is counted, and then the proportion of this number in the total number of Chinese medicine samples of the same Chinese medicine is calculated. When the proportion exceeds the preset proportion threshold, the Chinese medicine is marked as a special Chinese medicine. The residual is the difference between the organic acid content actually observed in the Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model;

[0155] The special content analysis module is used to construct a polynomial regression model based on the organic acid content of special Chinese medicine and the reflectivity of the corresponding wavelength, so as to predict the organic acid content of special Chinese medicine.

[0156] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0157] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. Those skilled in the art may appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein may be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed by hardware or software methods depends on the specific application and design constraints of the technical solution.

[0158] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0159] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, characterized in that: The specific steps include: Step 1: Obtain the content of organic acids in each TCM sample to be measured, use hyperspectral imaging technology, focus on the 400 to 2500 nm band, obtain the reflectivity of different wavelengths in each TCM sample, and construct a TCM data set with different wavelengths and corresponding reflectivity in each TCM; Step 2: Calculate the first-order and second-order derivatives of each wavelength, screen the wavelengths that may be associated with organic acids based on the derivative results, establish an organic acid-related data set using the wavelengths that may be associated with organic acids and the corresponding reflectance, and group the data according to the name of the traditional Chinese medicine; Step 3: Extract data from the organic acid-related data set, use the wavelengths that may be associated with organic acids and the corresponding reflectance to build a polynomial regression model, and select the polynomial regression model with the largest correlation value through the correlation coefficient formula to predict the organic acid content of traditional Chinese medicine; Step 4: Calculate the residuals of each Chinese medicine sample. For each Chinese medicine individual, count the number of Chinese medicine samples whose residuals exceed the preset residual threshold, and then calculate the proportion of this number in the total number of Chinese medicine samples of the same Chinese medicine. When the proportion exceeds the preset proportion threshold, mark this Chinese medicine as a special Chinese medicine. The residual is the difference between the organic acid content actually observed in the Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model. Step 5: A polynomial regression model is constructed separately for the organic acid content of the special Chinese medicine and the reflectance of the corresponding wavelength to predict the organic acid content of the special Chinese medicine.

2. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 1, obtaining the content of organic acid in each Chinese medicine sample to be determined includes the following steps: Grind the Chinese medicine sample into fine powder, add distilled water to fully soak the powder and stir for 30 minutes, remove solid impurities in the extract with filter paper, add distilled water, and titrate with 0.1 mol / L sodium hydroxide solution. When the solution color changes from colorless to pink and does not fade within half a minute, record the volume of sodium hydroxide solution consumed at this time, calculate the amount of sodium hydroxide consumed by the formula, that is, the amount of organic acid, and calculate the content of organic acid in the Chinese medicine sample.

3. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 2, characterized in that: Calculating the content of organic acids in a Chinese medicine sample by the amount of sodium hydroxide consumed includes the following steps: The amount of sodium hydroxide consumed is calculated using the formula n = c·V, where n represents the amount of sodium hydroxide, c represents the molar concentration of the sodium hydroxide solution, which is 0.1 mol / L, and V represents the volume of the sodium hydroxide solution added to the extract. The content of organic acids in the traditional Chinese medicine samples was calculated, where w represented the content of organic acids in the traditional Chinese medicine samples, M represented the molar mass of organic acids, and m represented the mass of traditional Chinese medicine samples.

4. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 1, obtaining the reflectance of each Chinese medicine sample at different wavelengths includes the following steps: Using hyperspectral imaging technology, the spectral range is pre-set to 400-2500nm and the resolution is set to 10nm. A hyperspectral image data cube is constructed. The reflected light signal intensity of each pixel point on the surface of the traditional Chinese medicine at each wavelength is directly obtained through the spectroscopic component and the detector. The average value of the reflected light signal intensity of all pixel points is calculated to construct the reflectance formula: Where R is the reflectivity, I yb It is expressed as the reflected light signal intensity of the standard white plate under the same lighting conditions and wavelength, I bb It is expressed as the reflected light signal intensity of the Chinese medicine sample at a certain wavelength.

5. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 1, the method of constructing a Chinese medicine dataset by combining different wavelengths and corresponding reflectances of each Chinese medicine is as follows: For each Chinese medicine sample, the reflectance data at each wavelength is obtained according to the pre-set spectral range of 400nm to 2500nm and 10nm resolution. All wavelength data corresponding to each Chinese medicine sample are taken as a row, and the values ​​of each column in the row are the reflectance of the Chinese medicine sample at the corresponding wavelength. Different rows of Chinese medicine sample data are arranged in sequence to construct a two-dimensional Chinese medicine data set, in which rows represent different Chinese medicine sample individuals, and columns represent different wavelengths and their corresponding reflectance data.

6. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 2, the first-order and second-order derivatives of each wavelength are calculated, and the wavelengths that may be associated with organic acids are screened based on the derivative results, including the following steps: Extract data from the data collection table and pre-set the reflectivity sequence to R, R = {r 400 , r 410 ,…,r 2400 , r 2500 }, where r refers to the reflectivity at each wavelength, and the wavelength sequence is set to λ, λ={λ 400 ,λ 410 , …, λ 2400 ,λ 2500 }, indicating the wavelength value starting from 400nm and extending to 2500nm with an interval of 10nm, and setting the wavelength interval to Δλ. Since the resolution is 10nm, Δλ=10; Using the central difference method, calculate the first derivative sequence of the reflectivity sequence R with respect to the wavelength λ: Where i represents the wavelength, i = 400, 410, ..., 2500, R ′ (λ i ) represents the first-order derivative sequence of the reflectivity sequence R with respect to the wavelength λ, that is, at the wavelength λ i The rate of change of reflectivity at wavelength λ = 400 and 2500 is calculated by the forward difference method: In the formula, R ′ (λ 400 ) is expressed as the first-order derivative of the reflectivity series R with respect to the wavelength of 400nm, R ′ (λ 400 ) is expressed as the first-order derivative of the reflectivity series R with respect to the wavelength of 2500nm, Based on the first-order derivative sequence, the second-order derivative is calculated using the central difference method: In the formula, R″(λ i ) represents the reflectivity sequence R with respect to wavelength λ at wavelength value λ i The value of the second-order derivative sequence at , The calculation of the second-order derivative of the boundary point is similar to the processing of the first-order derivative boundary point, requiring a special one-sided difference formula: In the formula, R″(λ 400 ) represents the reflectivity sequence R with respect to the wavelength value λ 400 The value of the second-order derivative sequence at , R″(λ 2500 ) represents the reflectivity sequence R with respect to the wavelength value λ 2500 The value of the second-order derivative sequence at , The position where the first-order derivative is zero and the second-order derivative is less than zero is defined as the absorption peak position, and the wavelength corresponding to the absorption peak position is recorded. At a certain wavelength, the number of Chinese medicine samples corresponding to its absorption peak reaches 90% or more of the total number of Chinese medicine samples. This wavelength can be preliminarily identified as a wavelength that may be associated with organic acids.

7. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 3, constructing a polynomial regression model and selecting the model with the largest correlation value to predict the organic acid content of the traditional Chinese medicine includes the following steps: First, the data of the same traditional Chinese medicine in the organic acid association data set were extracted respectively, and the reflectance corresponding to the wavelengths that were initially identified as possibly associated with organic acids was used as the independent variable, and the corresponding organic acid content data of the same traditional Chinese medicine was used as the dependent variable to construct a quadratic polynomial regression model: Where x represents the wavelength reflectance data, is expressed as the predicted organic acid content data, β0, β1 and β2 are expressed as coefficients, The parameters β0, β1 and β2 in the model are determined by the least squares method, that is, the parameters β0, β1 and β2 in the model are determined by minimizing the sum of squared errors: In the formula, SSE means minimizing the sum of squared errors, n is the number of TCM samples, and x a is the reflectance of the ath TCM sample at the wavelength that is preliminarily identified as possibly associated with organic acids, y a is the organic acid content of the a-th Chinese medicine sample, a=1, 2, …, z, where z is the number of Chinese medicines collected, and z is a positive integer ≥ 50. The partial derivatives with respect to β0, β1 and β2 are calculated respectively, and the partial derivatives are set equal to zero to solve the quadratic polynomial regression model; For β0: In the formula, It means to find the partial derivative of β0; For β1: In the formula, It means to find the partial derivative of β1; For β2: In the formula, It means to find the partial derivative of β2; By solving the above system of three partial derivative equations, we can obtain the values ​​of β0, β1 and β2; Then, the correlation coefficient between the reflectance data and the organic acid content data under the model was calculated; The correlation coefficient calculation formula is as follows: In the formula, r represents the correlation coefficient between the reflectance data and the organic acid content data, is the mean value of the reflectance data corresponding to the wavelength that is preliminarily identified as possibly associated with organic acids. is the mean of the organic acid content data, Repeat the above process of building the model and calculating the correlation coefficient to obtain multiple different quadratic polynomial regression models and their corresponding correlation coefficients; Finally, these correlation coefficients were compared, and the wavelength with the highest correlation coefficient value r was selected and determined as the specific wavelength related to the content of organic acids in traditional Chinese medicine. The polynomial regression model with the largest correlation value was selected to predict the organic acid content of traditional Chinese medicine.

8. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 4, the methods marked as special Chinese medicine are: For each Chinese medicine sample, the residual is calculated separately, and the residual calculation formula is: In the formula, represents the organic acid content calculated by the quadratic polynomial regression model, e a It represents the difference between the organic acid content actually observed in each Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model. The residual threshold is pre-set to e max , the ratio threshold is q max , for each Chinese medicine, the residual e is counted a Greater than or equal to the residual threshold e max The number of Chinese medicine samples is calculated, and then the proportion of the number of Chinese medicine samples that meet the above conditions in the total number of Chinese medicine samples of this type of Chinese medicine is calculated. If it exceeds the proportion threshold q max , marking this Chinese medicine as a special Chinese medicine.

9. The method for predicting the organic acid content of traditional Chinese medicine based on hyperspectral imaging according to claim 1, characterized in that: In step 5, the method of constructing a polynomial regression model alone is: For the types marked as special Chinese medicines, the extracted data include: name of the Chinese medicine, actual organic acid content, reflectivity at specific wavelengths, The specific wavelength reflectance data x is used as the independent variable, and the corresponding organic acid content data y is used as the dependent variable to construct a quadratic polynomial regression model: The parameters β0, β1 and β2 in the model are determined by the least squares method, that is, the parameters β0, β1 and β2 in the model are determined by minimizing the sum of squared errors: Calculate the partial derivatives with respect to β0, β1 and β2 respectively, and set the partial derivatives equal to zero to solve the quadratic polynomial regression model.

10. A system for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging, characterized in that: The system is used to execute a method for predicting the content of organic acids in traditional Chinese medicine based on hyperspectral imaging as described in any one of claims 1 to 9: The data set acquisition module is used to obtain the content of organic acids in each Chinese medicine sample to be measured, and use hyperspectral imaging technology to focus on the 400 to 2500nm band to obtain the reflectivity of different wavelengths in each Chinese medicine sample, and construct a Chinese medicine data set with different wavelengths and corresponding reflectivity in each Chinese medicine; The data grouping module is used to calculate the first-order and second-order derivatives of each wavelength, screen the wavelengths that may be associated with organic acids based on the derivative results, establish an organic acid-related data set using the wavelengths that may be associated with organic acids and the corresponding reflectance, and group the data according to the name of the traditional Chinese medicine; The linear regression module is used to extract data from the organic acid-related data set, build a polynomial regression model using the wavelengths that may be associated with organic acids and the corresponding reflectance, and select the polynomial regression model with the largest correlation value through the correlation coefficient formula to predict the organic acid content of traditional Chinese medicine; The difference analysis module is used to calculate the residuals of each Chinese medicine sample. For each Chinese medicine individual, the number of Chinese medicine samples whose residuals exceed the preset residual threshold is counted, and then the proportion of this number in the total number of Chinese medicine samples of the same Chinese medicine is calculated. When the proportion exceeds the preset proportion threshold, the Chinese medicine is marked as a special Chinese medicine. The residual is the difference between the organic acid content actually observed in the Chinese medicine sample and the organic acid content calculated by the quadratic polynomial regression model; The special content analysis module is used to construct a polynomial regression model based on the organic acid content of special Chinese medicine and the reflectivity of the corresponding wavelength, so as to predict the organic acid content of special Chinese medicine.

Images