Grape aspergillosis spore detection method based on multispectral scattering

Abstract

The invention discloses a grape aspergillosis spore detection method based on multispectral scattering, which relates to the technical field of agricultural product disease detection, and comprises the steps of sample standardization processing, multidimensional scattering data acquisition, scattering characteristic parameter extraction and fusion, fingerprint identification library construction, and qualitative and quantitative model training. And rapid and accurate identification and concentration quantification of the grape aspergillosis spores are realized. The core of the method is to fuse multispectral information and particle morphological characteristics, and through characteristic wavelength screening and machine learning model construction, the problems of long time consumption, insufficient specificity and incapability of quantification of an existing detection method are solved. The method is high in detection efficiency, strong in specificity and high in quantification precision, and can be widely applied to disease early warning and prevention and control in grape planting and postpartum storage processes.

Description

Technical Field

[0001] This invention relates to the field of agricultural product disease detection technology, and more specifically to a method for detecting grape aspergillosis spores based on multispectral scattering. Background Technology

[0002] Currently, grape aspergillosis is a significant disease caused by fungi of the genus Aspergillus. Its spores multiply rapidly under suitable temperature and humidity conditions, leading to grape rot and spoilage, severely impacting grape yield and quality. Current detection methods for grape aspergillosis mainly rely on traditional culture methods, microscopic observation, and chemical detection. However, these methods have significant drawbacks: traditional culture methods take 24-48 hours, failing to meet the need for rapid detection; microscopic observation relies on manual identification, is inefficient, and is easily influenced by operator experience, making it difficult to distinguish morphologically similar aspergillosis spores from other microparticles (such as dust or yeast cells); chemical detection methods pose a risk of reagent contamination and cannot achieve quantitative spore counting.

[0003] Existing spectroscopic detection technologies (such as near-infrared spectroscopy and Raman spectroscopy) achieve detection by analyzing the spectral characteristics of substances. However, scattering detection with a single wavelength or a limited wavelength is difficult to capture the subtle differences between spores and interfering particles. Since the particle size and refractive index of Aspergillus spores and some interfering microparticles are similar, their scattering characteristics highly overlap at a single wavelength, resulting in insufficient detection specificity.

[0004] Therefore, how to integrate multi-dimensional spectral information and morphological features to construct a specific recognition model is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0005] In view of this, the present invention provides a method for detecting grape aspergillosis spores based on multispectral scattering, in order to solve the problems existing in the background art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A method for detecting Aspergillus spores of grapevines based on multispectral scattering, comprising: S1. Obtain a sample containing grape aspergillosis spores and interfering microparticles, and prepare a standardized sample after purification and dispersion. S2. Using a multispectral scattering acquisition system, multiple characteristic wavelengths are selected, and the scattered light intensity signals of the standardized sample at each characteristic wavelength are acquired within a set scattering angle range to obtain multidimensional scattering data. S3. Based on the multi-dimensional scattering data, extract scattering feature parameters related to the particle morphology characteristics in the standardized samples; S4. Standardize the scattering characteristic parameters under each characteristic wavelength and fuse them to obtain the multispectral scattering fingerprint characteristics of the sample. S5. Collect standard samples of grape aspergillosis spores and standard samples of interfering microparticles. Extract the multispectral scattering fingerprint features of each standard sample according to S2-S4. After labeling the category, construct a fingerprint recognition database. S6. Based on the fingerprint recognition database, a qualitative detection model for distinguishing between grape aspergillosis spores and interfering microparticles is trained; S7. Based on the correlation between the multispectral scattering fingerprint characteristics and concentration of standard samples of grape aspergillosis spores, a quantitative identification model was established; S8. Process the sample to be tested according to S2-S4 to obtain its multispectral scattering fingerprint characteristics, and input them into the qualitative detection model to obtain the category determination result; if it is determined to be grape aspergillosis spores, obtain its concentration through the quantitative identification model.

[0007] Optionally, S1 further includes controlling sample homogeneity through the particle size distribution variation coefficient, wherein the particle size distribution variation coefficient satisfies:

[0008] Where CV is the coefficient of variation of particle size distribution. The standard deviation of particle size. The average particle size; The concentration of the suspension was calibrated using a hemocytometer. The concentration calibration formula is as follows:

[0009] In the formula, C is the actual concentration, and N is the total number of particles in the counting cell. The total volume of the counting plate To dilute the volume, This represents the actual number of cells counted.

[0010] Optionally, S2 specifically includes: Based on Mie scattering theory, a multispectral scattering acquisition system is employed. Multiple characteristic wavelengths are selected using a characteristic wavelength discrimination formula. Within a set scattering angle range, the scattered light intensity signals of the standardized sample at each characteristic wavelength are acquired, resulting in a multidimensional scattering data matrix. The specific characteristic wavelength discrimination formula is as follows:

[0011] In the formula, wavelength Discrimination index, The scattering characteristic entropy of grape aspergillosis spores. The scattering characteristic entropy of the interfering microparticles.

[0012] Optionally, the scattering characteristic parameters include the position of the main peak, the number of scattering peaks, the area of ​​the scattered light intensity integral, the full width at half maximum (FWHM), and the ratio of the intensity of the main peak to that of the secondary peak.

[0013] Optionally, S5 specifically includes: based on statistical learning theory, collecting standard samples of grape aspergillus spores covering different concentrations and growth stages, as well as standard samples of interfering microparticles from different sources; extracting fingerprint feature vectors of each standard sample according to S2-S4; labeling with category and concentration tags; and constructing a fingerprint recognition library that meets the requirements of concentration coverage and interference type coverage. Optionally, S6 specifically includes: based on machine learning classification theory, using multispectral scattering fingerprint features in the fingerprint recognition library as input and category tags as output, constructing a qualitative detection model. The qualitative detection model learns the mapping relationship between multispectral scattering fingerprint features and category tags to form an optimal decision boundary to distinguish between grape aspergillus spores and interfering microparticles. Optionally, the loss function of the qualitative detection model is specifically:

[0014] In the formula, L ( f ) represents the total loss function; For prediction functions; Here, N is the classification loss function; N is the number of spore samples; and M is the number of interfering samples. Category labels; Let be the fingerprint feature vector of the s-th standard sample; The regularization coefficient is used. This is a complexity penalty term.

[0015] Optionally, S7 specifically includes: based on the linear correlation principle extended from Lambert-Beer's law, selecting the optimal characteristic wavelength through the Pearson correlation coefficient formula, and using the multispectral scattering fingerprint characteristics under the optimal characteristic wavelength as the independent variable and the spore concentration as the dependent variable, establishing a quantitative mapping model between the multispectral scattering fingerprint characteristics and the spore concentration to realize the calculation of the spore concentration.

[0016] Optionally, the selection of the optimal feature wavelength specifically involves: Calculate the correlation coefficient between scattering characteristics and concentration at each wavelength, and select the wavelength with the strongest correlation:

[0017] In the formula, r ( λ S(λ) is the Pearson correlation coefficient between the integral area S(λ) at wavelength λ and the concentration C; S(λ,t) is the integral area of ​​the t-th concentration sample at wavelength λ. This is the mean area of ​​the integral over wavelength λ for all concentration samples. The actual concentration of the t-th sample; The actual concentration mean of all concentration samples; T is the number of concentration samples; The optimal wavelength is then calculated using the following formula: .

[0018] As can be seen from the above technical solution, compared with the prior art, the present invention discloses a method for detecting grape aspergillosis spores based on multispectral scattering, which has the following beneficial effects: 1. This invention does not require a complex sample culture process, the entire detection process is less time-consuming and much faster than traditional culture methods, and the detection efficiency is high, meeting the needs of rapid detection; 2. This invention captures the subtle differences between grape aspergillosis spores and interfering microparticles through multispectral information fusion and characteristic wavelength screening, effectively solving the problem of insufficient specificity of single-wavelength detection and achieving high qualitative accuracy. 3. Based on the Lambert-Beer law and Pearson correlation coefficient, this invention selects the optimal characteristic wavelength and establishes a quantitative model with high accuracy and small relative error, which can accurately obtain spore concentration and provide data support for disease early warning. 4. This invention does not require the use of chemical reagents, thus avoiding the risk of reagent contamination. Furthermore, the operation process is highly automated, has little dependence on the experience of operators, and is easy to promote and apply.

[0019] This invention can be widely applied to the detection of grape aspergillosis spores in grape planting bases, wineries, fruit and vegetable storage facilities, etc., providing technical support for early warning and precise control of grape aspergillosis, and has important practical application value. Attached Figure Description

[0020] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0021] Figure 1 This is a schematic diagram of the method flow provided by the present invention. Detailed Implementation

[0022] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0023] This invention discloses a method for detecting Aspergillus spores of grapevine based on multispectral scattering, such as... Figure 1 As shown, it includes: S1. Obtain a sample containing grape aspergillosis spores and interfering microparticles, and prepare a standardized sample after purification and dispersion; specifically: Sample acquisition: Air samples and grape fruit surface swab samples were collected from the vineyard and storage environment. The samples contained grape aspergillosis spores and interfering microparticles such as dust, yeast cells, and Penicillium spores.

[0024] Purification process: Large particulate impurities in the sample are removed by membrane filtration, and spores and microparticles are purified by centrifugation.

[0025] Dispersion treatment: The purified sample is added to sterile physiological saline and dispersed using an ultrasonic disperser to avoid particle aggregation.

[0026] Uniformity control: The particle size distribution is detected by a laser particle size analyzer, and the particle size distribution variation coefficient (CV) is calculated to ensure that the CV is less than or equal to 0.2. If the CV is not met, the dispersion process is repeated.

[0027] Concentration calibration: The concentration of the suspension was calibrated using a hemocytometer. Three counting cells were selected for counting, and the average value was taken as the total number of particles N in the counting cells. The actual concentration was then calculated by substituting the average value into the concentration calibration formula.

[0028] S2. A multispectral scattering acquisition system is used to select multiple characteristic wavelengths and collect the scattered light intensity signals of the standardized sample at each characteristic wavelength within a set scattering angle range to obtain multidimensional scattering data. The multispectral scattering acquisition system includes a xenon lamp light source, a monochromator, and a photodetector array. The scattering angle detection range is set to 5°-120°.

[0029] Feature wavelength screening: Select wavelengths in the range of 400nm-1000nm, select candidate wavelengths at 10nm intervals, calculate the discrimination index of each candidate wavelength, and select the top 5 wavelengths with the highest discrimination index as feature wavelengths, namely 520nm, 630nm, 750nm, 880nm, and 940nm.

[0030] Scattering data acquisition: Standardized samples are injected into quartz cuvettes and placed in the detection station of the multispectral scattering acquisition system. Scattered light intensity signals within the scattering angle range of 5°-120° are acquired at 5 set characteristic wavelengths. Each scattering angle is acquired 3 times, and the average value is taken as the light intensity data at that angle, forming a 5×116 (number of characteristic wavelengths × number of scattering angles) multidimensional scattering data matrix.

[0031] S3. Based on the multi-dimensional scattering data, extract scattering feature parameters related to the particle morphology characteristics in the standardized samples; Specifically, based on a multi-dimensional scattering data matrix, a peak detection algorithm is used to extract scattering characteristic parameters at each characteristic wavelength: Main peak position: the scattering angle corresponding to the maximum scattered light intensity;

[0032] In the formula, The location of the main peak; A is the peak height coefficient; B is the peak width coefficient; B is the background light intensity. Scattering angle fitted using the least squares method and corresponding light intensity Solve .

[0033] Number of scattering peaks: The number of peaks with light intensity values ​​greater than the threshold; Scattered light intensity integral area: the integral value of the light intensity signal within the scattering angle range of 5°-120°; Scattered light intensity integral area: reflects the total light scattering ability of particles, and is positively correlated with particle concentration and particle size.

[0034] Discretization computation:

[0035] In the formula, wavelength The area of ​​the scattered light intensity integral; The step size for sampling the scattering angle; , The minimum and maximum values ​​for the scattering angle are collected. Let be the light intensity at the j-th scattering angle.

[0036] Half-width at half height (HWHM): The range of scattering angles corresponding to half the height of the main peak; The ratio of the light intensity of the main peak to that of the secondary peak is the ratio of the maximum light intensity of the main peak to the maximum light intensity of the secondary peak.

[0037]

[0038] In the formula, The ratio of light intensity; The intensity of the scattered light corresponding to the secondary peak is detected by the sliding window method.

[0039] S4. Standardize the scattering characteristic parameters under each characteristic wavelength and fuse them to obtain the multispectral scattering fingerprint characteristics of the sample. Specifically, standardization processing: The Z-score standardization method is used to process each feature parameter, and the formula is as follows: (where x is the original parameter value,) This is the mean of the parameter. (This is the standard deviation of the parameter).

[0040] Fingerprint feature fusion: Five standardized feature parameters under five feature wavelengths are sequentially spliced ​​together to form a 25-dimensional multispectral scattering fingerprint feature vector.

[0041] S5. Collect standard samples of grape aspergillosis spores and standard samples of interfering microparticles. Extract the multispectral scattering fingerprint features of each standard sample according to S2-S4. After labeling the category, construct a fingerprint recognition database. Specifically, standard sample collection: Standard samples of Aspergillus spores for grapevine disease: Aspergillus spores at different growth stages (spore germination, maturity, and senescence) were cultured to prepare samples with a concentration of 10³⁻¹. A series of concentration samples per mL, totaling 30 samples; Interference microparticle standard samples: Interference microparticles from different sources such as dust, yeast cells, Penicillium spores, and Mucor spores were collected and prepared at a concentration of 10³⁻¹. Samples per mL, with 5 samples for each type of interference, for a total of 20 samples.

[0042] Fingerprint recognition database construction: Extract multispectral scattering fingerprint feature vectors from the above 50 standard samples, label them with category labels (grape aspergillosis spores are labeled as 1, interfering microparticles are labeled as 0) and concentration labels, and construct the fingerprint recognition database.

[0043] S6. Based on the fingerprint recognition database, a qualitative detection model for distinguishing between grape aspergillosis spores and interfering microparticles is trained. The qualitative detection model training uses a support vector machine (SVM) as the classification model, with a 25-dimensional fingerprint feature vector from the fingerprint recognition database as input and the category label as output, and sets a regularization coefficient. The classification loss function adopts the hinge loss function, the complexity penalty term adopts L2 regularization, and the model parameters are optimized through cross-validation to obtain a qualitative detection model.

[0044] S7. Based on the correlation between the multispectral scattering fingerprint characteristics and concentration of grape aspergillosis spore standard samples, a quantitative identification model was established. The quantitative identification model was established by calculating the Pearson correlation coefficient between the integrated area of ​​scattered light intensity and spore concentration at each characteristic wavelength. The correlation coefficient r(750nm) = 0.982 at 750nm wavelength was found to be the optimal characteristic wavelength. A linear regression model was established with the integrated area of ​​scattered light intensity at 750nm wavelength as the independent variable and spore concentration as the dependent variable. , where a and b are model fitting coefficients, and S is the normalized scattering characteristic at the optimal wavelength.

[0045] S8. Process the sample to be tested according to S2-S4 to obtain its multispectral scattering fingerprint characteristics, and input them into the qualitative detection model to obtain the category determination result; if it is determined to be grape aspergillosis spores, obtain its concentration through the quantitative identification model.

[0046] In a specific embodiment, S1 further includes controlling sample homogeneity through the particle size distribution variation coefficient, wherein the particle size distribution variation coefficient satisfies:

[0047] Where CV is the coefficient of variation of particle size distribution. The standard deviation of particle size. The average particle size; The concentration of the suspension was calibrated using a hemocytometer. The concentration calibration formula is as follows:

[0048] In the formula, C is the actual concentration, and N is the total number of particles in the counting cell. The total volume of the counting plate To dilute the volume, This represents the actual number of cells counted.

[0049] In one specific embodiment, S2 specifically includes: Based on Mie scattering theory, a multispectral scattering acquisition system is employed. Multiple characteristic wavelengths are selected using a characteristic wavelength discrimination formula. Within a set scattering angle range, the scattered light intensity signals of the standardized sample at each characteristic wavelength are acquired, resulting in a multidimensional scattering data matrix. The specific characteristic wavelength discrimination formula is as follows:

[0050] In the formula, wavelength Discrimination index, The scattering characteristic entropy of grape aspergillosis spores. The scattering characteristic entropy of the interfering microparticles.

[0051] In one specific embodiment, S5 specifically includes: based on statistical learning theory, collecting standard samples of grape aspergillosis spores covering different concentrations and growth stages, as well as standard samples of interfering microparticles from different sources; extracting fingerprint feature vectors of each standard sample according to S2-S4; labeling with category and concentration tags; and constructing a fingerprint recognition library that meets the requirements of concentration coverage and interference type coverage. In one specific embodiment, S6 specifically includes: based on machine learning classification theory, using multispectral scattering fingerprint features from the fingerprint recognition library as input and category tags as output, constructing a qualitative detection model. This qualitative detection model learns the mapping relationship between multispectral scattering fingerprint features and category tags to form an optimal decision boundary to distinguish between grape aspergillosis spores and interfering microparticles. In one specific embodiment, the loss function of the qualitative detection model is specifically:

[0052] In the formula, L ( f ) represents the total loss function; For prediction functions; Here, N is the classification loss function; N is the number of spore samples; and M is the number of interfering samples. Category labels; Let be the fingerprint feature vector of the s-th standard sample; The regularization coefficient is used. This is a complexity penalty term.

[0053] In a specific embodiment, S7 specifically includes: based on the linear correlation principle extended from Lambert-Beer's law, selecting the optimal characteristic wavelength through the Pearson correlation coefficient formula, and using the multispectral scattering fingerprint characteristics under the optimal characteristic wavelength as the independent variable and the spore concentration as the dependent variable, establishing a quantitative mapping model between the multispectral scattering fingerprint characteristics and the spore concentration to realize the calculation of the spore concentration.

[0054] In a specific embodiment, the selection of the optimal feature wavelength is specifically as follows: Calculate the correlation coefficient between scattering characteristics and concentration at each wavelength, and select the wavelength with the strongest correlation:

[0055] In the formula, r ( λ S(λ) is the Pearson correlation coefficient between the integral area S(λ) at wavelength λ and the concentration C; S(λ,t) is the integral area of ​​the t-th concentration sample at wavelength λ. This is the mean area of ​​the integral over wavelength λ for all concentration samples. The actual concentration of the t-th sample; The actual concentration mean of all concentration samples; T is the number of concentration samples; The optimal wavelength is then calculated using the following formula: .

[0056] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since they correspond to the methods disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to the method section.

[0057] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for detecting Aspergillus spores of grapevine based on multispectral scattering, characterized in that, include: S1. Obtain a sample containing grape aspergillosis spores and interfering microparticles, and prepare a standardized sample after purification and dispersion. S2. Using a multispectral scattering acquisition system, multiple characteristic wavelengths are selected, and the scattered light intensity signals of the standardized sample at each characteristic wavelength are acquired within a set scattering angle range to obtain multidimensional scattering data. S3. Based on the multi-dimensional scattering data, extract scattering feature parameters related to the particle morphology characteristics in the standardized samples; S4. Standardize the scattering characteristic parameters under each characteristic wavelength and fuse them to obtain the multispectral scattering fingerprint characteristics of the sample. S5. Collect standard samples of grape aspergillosis spores and standard samples of interfering microparticles. Extract the multispectral scattering fingerprint features of each standard sample according to S2-S4, label the category, and construct a fingerprint recognition database. S6. Based on the fingerprint recognition database, a qualitative detection model for distinguishing between grape aspergillosis spores and interfering microparticles is trained. S7. Based on the correlation between the multispectral scattering fingerprint characteristics and concentration of standard samples of grape aspergillosis spores, a quantitative identification model was established; S8. Process the sample to be tested according to S2-S4 to obtain its multispectral scattering fingerprint characteristics, and input them into the qualitative detection model to obtain the category determination result; if it is determined to be grape aspergillosis spores, obtain its concentration through the quantitative identification model.

2. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 1, characterized in that, S1 further includes controlling sample homogeneity through the particle size distribution variation coefficient, wherein the particle size distribution variation coefficient satisfies: Where CV is the coefficient of variation of particle size distribution. The standard deviation of particle size. The average particle size; The concentration of the suspension was calibrated using a hemocytometer. The concentration calibration formula is as follows: In the formula, C is the actual concentration, and N is the total number of particles in the counting cell. The total volume of the counting plate To dilute the volume, This represents the actual number of cells counted.

3. The method for detecting Aspergillus spores of grapevines based on multispectral scattering according to claim 1, characterized in that, S2 specifically includes: Based on Mie scattering theory, a multispectral scattering acquisition system is employed. Multiple characteristic wavelengths are selected using a characteristic wavelength discrimination formula. Within a set scattering angle range, the scattered light intensity signals of the standardized sample at each characteristic wavelength are acquired, resulting in a multidimensional scattering data matrix. The specific characteristic wavelength discrimination formula is as follows: In the formula, wavelength Discrimination index, The scattering characteristic entropy of grape aspergillosis spores. The scattering characteristic entropy of the interfering microparticles.

4. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 1, characterized in that, The scattering characteristic parameters include the position of the main peak, the number of scattering peaks, the area of ​​the scattered light intensity integral, the half-width at half-maximum (WHM), and the ratio of the intensity of the main peak to that of the secondary peak.

5. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 1, characterized in that, S5 specifically includes: based on statistical learning theory, collecting standard samples of grape aspergillus spores covering different concentrations and different growth stages, as well as standard samples of interfering microparticles from different sources; extracting fingerprint feature vectors of each standard sample according to S2-S4; labeling them with category and concentration labels; and constructing a fingerprint recognition library that meets the requirements of concentration coverage and interference type coverage.

6. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 1, characterized in that, Specifically, S6 involves: based on machine learning classification theory, using multispectral scattering fingerprint features from the fingerprint recognition database as input and category labels as output, constructing a qualitative detection model. The qualitative detection model learns the mapping relationship between multispectral scattering fingerprint features and category labels to form an optimal decision boundary to distinguish between grape aspergillosis spores and interfering microparticles.

7. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 6, characterized in that, The loss function of the qualitative detection model is specifically as follows: In the formula, L ( f ) represents the total loss function; For prediction functions; Here, N is the classification loss function; N is the number of spore samples; and M is the number of interfering samples. Category labels; Let be the fingerprint feature vector of the s-th standard sample; The regularization coefficient is used. This is a complexity penalty term.

8. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 1, characterized in that, Specifically, S7 includes: based on the linear correlation principle extended from Lambert-Beer's law, selecting the optimal characteristic wavelength through the Pearson correlation coefficient formula, and establishing a quantitative mapping model between multispectral scattering fingerprint characteristics and spore concentration with the multispectral scattering fingerprint characteristics under the optimal characteristic wavelength as the independent variable and spore concentration as the dependent variable, thereby realizing the calculation of spore concentration.

9. The method for detecting Aspergillus spores of grapevine based on multispectral scattering according to claim 8, characterized in that, The specific steps for selecting the optimal feature wavelength are as follows: Calculate the correlation coefficient between scattering characteristics and concentration at each wavelength, and select the wavelength with the strongest correlation: In the formula, r ( λ S(λ) is the Pearson correlation coefficient between the integral area S(λ) at wavelength λ and the concentration C; S(λ,t) is the integral area of ​​the t-th concentration sample at wavelength λ. This is the mean area of ​​the integral over wavelength λ for all concentration samples. The actual concentration of the t-th sample; The actual concentration mean of all concentration samples; T is the number of concentration samples; The optimal wavelength is then calculated using the following formula: 。

Images