An infrared spectroscopy chemometric analysis method and system

Abstract

The application discloses an infrared spectrum chemometrics analysis method and system, and relates to the technical field of infrared spectrum analysis. The method comprises the following steps: acquiring infrared spectrum images with known spectrum characteristic parameters, pre-processing the infrared spectrum images to obtain a training sample image set; based on the training sample image set, a neural network prediction model is established, the infrared spectrum images are taken as input, the spectrum characteristic parameters are taken as labels, and a spectrum characteristic parameter extraction model is trained and obtained; an infrared spectrum image of a to-be-detected object is collected, is input into the trained model after pre-processing, and spectrum characteristic parameters are output; meanwhile, environmental parameters are collected; the spectrum characteristic parameters are corrected based on the environmental parameters, and the spectrum characteristic parameters are further corrected to obtain accurate values in combination with physical characteristic parameters of the to-be-detected object; and chemometrics analysis of properties and contents of the to-be-detected object is completed based on the accurate spectrum characteristic parameters. The real values of the spectrum characteristic parameters are accurately simulated, and the accuracy of analysis results is significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of infrared spectrum analysis, in particular to an infrared spectrum chemometrics analysis method and system. BACKGROUND

[0002] As a rapid and non-destructive analysis tool, infrared spectrum analysis is widely used in the fields of chemistry, material science, biological medicine, etc., for studying the molecular structure and chemical composition of substances. By analyzing spectral characteristic parameters (such as peak position, peak intensity and integral area), researchers can obtain the chemical properties and physical characteristics of the sample. However, in practical applications, infrared spectrum analysis is limited by many technical problems, making it difficult to accurately analyze.

[0003] Firstly, infrared spectrum signals are more susceptible to experimental conditions and environmental factors, such as environmental temperature, humidity, and light intensity during the collection process. These parameters can significantly affect the spectrum signal. For example, changes in environmental temperature can cause the response of the infrared detector to be unstable, affecting the peak position and intensity of the spectrum; differences in light intensity can introduce additional background noise or baseline drift, interfering with the authenticity of the spectrum signal. The interference of these external environmental factors makes it difficult for traditional spectral analysis methods to directly obtain accurate spectral characteristic parameters, and additional correction processing is required. However, existing correction methods are mostly simple linear models that do not fully consider the complex nonlinear relationship between environmental parameters and spectral characteristics, resulting in low correction accuracy.

[0004] Secondly, the characteristic parameters of infrared spectrum are not only affected by environmental factors, but also significantly affected by the physical characteristics of the sample itself (such as particle size distribution, surface roughness and thickness). For example, the particle size of particulate matter will affect the scattering and transmission behavior of light, causing peak intensity deviation or peak position shift; changes in the surface roughness of the sample will enhance the diffuse reflection of light, weakening the intensity of the spectrum signal; differences in thickness will directly affect the absorption intensity of light, which may cause significant changes in the integral area of the spectrum. Existing infrared spectrum analysis methods usually ignore the influence of these sample physical characteristics or only use simple empirical formulas for correction, resulting in insufficient accuracy and robustness of the analysis results.

[0005] In summary, existing infrared spectrum analysis methods still have obvious shortcomings in dealing with environmental factor interference, sample physical characteristic influence and complex spectrum feature extraction. These problems limit the widespread application of infrared spectrum technology in high-precision analysis fields. Therefore, there is an urgent need for an infrared spectrum chemometrics analysis method that can comprehensively consider environmental parameters and sample physical characteristics to improve the accuracy of spectral characteristic parameters and the reliability of analysis results.

[0006] In the prior art, the publication number CN110909470B discloses an infrared spectrum chemometrics analysis system and method, which comprises: a model establishing module, which is used to establish a sample set according to a pre-stored sample set file in the system, and generate a quantitative model and a qualitative model corresponding to the sample property of the sample in the sample set, and bind the sample and the several quantitative models corresponding to the sample, and bind the sample set and the several qualitative models and quantitative models corresponding to the sample set; a model transmission module, which is used to generate a spectrum correction factor F for correcting the spectrum from the sub-spectrometer spectrum to the mother spectrometer spectrum, and evaluate the accuracy of F; an analysis prediction module, which transmits the unknown spectrum sample collected by the sub-spectrometer to the mother spectrometer through the spectrum correction factor, and predicts the sample property and the corresponding sample property value contained in the spectrum transmitted to the mother spectrometer through the qualitative model and the quantitative model. During the spectrum collection process, factors such as environmental temperature, humidity and light intensity may cause significant interference to the spectrum signal, but this scheme does not specially process the influence of such environmental parameters. Ignoring environmental factors may cause instability of the prediction result, especially in complex actual application scenarios. Therefore, the accuracy and effectiveness of the analysis system are reduced.

[0007] The above information disclosed in the background section is only intended to enhance the understanding of the background of the present disclosure, and thus it can include information that does not constitute the prior art known to those of ordinary skill in the art. SUMMARY

[0008] The purpose of the present application is to provide an infrared spectrum chemometrics analysis method and system to solve the problems raised in the background.

[0009] To achieve the above-mentioned purpose, the present application provides the following technical solutions:

[0010] An infrared spectrum chemometrics analysis method, the specific steps comprising:

[0011] Obtain several infrared spectrum images of known spectrum characteristic parameters, preprocess the collected infrared spectrum images to obtain a training sample image set, wherein the preprocessing includes baseline correction preprocessing and noise removal preprocessing, and the spectrum characteristic parameters include peak position, peak intensity and integral area;

[0012] Based on the obtained training sample image set, a neural network prediction model is established, the infrared spectrum images in the training sample image set are taken as the input of the neural network prediction model, and the corresponding spectrum characteristic parameters are taken as the label, the neural network prediction model is trained, and a spectrum characteristic parameter extraction model is obtained;

[0013] An infrared spectrum image of the to-be-detected article is acquired, the infrared spectrum image of the to-be-detected article is preprocessed, and then input into a spectrum feature parameter extraction model that has been trained, to output spectrum feature parameters of the to-be-detected article; meanwhile, environmental parameters when the infrared spectrum image of the to-be-detected article is acquired are collected, and the environmental parameters include the environmental temperature of the collection place and the illumination intensity of the collection place;

[0014] Based on the collected environmental parameters, the output spectrum feature parameters of the to-be-detected article are corrected to obtain environmental correction feature parameters, the physical characteristic parameters of the to-be-detected article are analyzed, the environmental correction feature parameters are corrected based on the obtained physical characteristic parameters to obtain accurate spectrum feature parameters of the to-be-detected article, and based on the obtained accurate spectrum feature parameters of the to-be-detected article, stoichiometric analysis of the properties and content of the to-be-detected article is completed, and the physical characteristic parameters of the to-be-detected article include the average particle size, surface roughness and thickness of the to-be-detected article.

[0015] Further, the collected infrared spectrum image is preprocessed to obtain a training sample image set, wherein the preprocessing includes baseline correction preprocessing and noise removal preprocessing, and a polynomial fitting method is selected for baseline correction, and the specific operation steps include: a polynomial is fitted by a least square method to obtain baseline parameters, and then the baseline signal is subtracted from the original spectrum data to obtain a corrected spectrum.

[0016] In the noise removal preprocessing, an ensemble empirical mode decomposition method is used, and the specific steps include: different amplitudes of noise are added to the original spectrum, and the original spectrum is decomposed for several times, the IMFs obtained by the several times of decomposition are averaged, the low-frequency IMF component is retained, and the high-frequency component as noise is filtered out.

[0017] The generation method of the training sample image set is that the preprocessed infrared spectrum image and the spectrum feature parameters corresponding to the image are one-to-one mapped to form a corresponding grid, and the formed grid is recorded as the training sample image set.

[0018] Further, based on the obtained training sample image set, a neural network prediction model is established, wherein a long short-term memory network model (LSTM model) is selected as a base model, an activation function and an optimization algorithm are selected, a Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0019]

[0020] In the formula, represents the Tanh function, and the independent variable represents the input weighted sum of neurons, that is, the result of the input from the previous layer after being weighted and summed by the neurons;

[0021] Meanwhile, the hyperparameters of the LSTM model are set, including the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch quantity, and the number of hidden layer neurons.

[0022] The number of network layers is set to a three-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch quantity is set to 256, and the number of hidden layer neurons is set to 32.

[0023] The trained spectral feature parameter extraction model inputs the preprocessed infrared spectrum image and outputs the spectral feature parameters in the infrared spectrum image, including the peak position, the peak intensity, and the integral area.

[0024] Further, based on the collected environmental parameters, the output spectral feature parameters of the to-be-detected object are corrected to obtain environmental correction feature parameters, wherein the peak position is corrected by the environmental temperature to obtain a peak position environmental correction value, and the formula for calculating the peak position environmental correction value is:

[0025]

[0026] In the formula, is the peak position environmental correction value of the infrared spectrum image of the to-be-detected object, is the peak position prediction value of the to-be-detected object output by the spectral feature parameter extraction model, is the environmental temperature of the collection site, is the reference temperature of the infrared spectrum image collection.

[0027] Further, the peak intensity and the integral area are corrected by the environmental temperature and the illumination intensity of the collection site to obtain a peak intensity environmental correction value and an integral area environmental correction value, and the formula for calculating the peak intensity environmental correction value is:

[0028]

[0029] In the formula, is the peak intensity environmental correction value, is the peak intensity prediction value of the to-be-detected object output by the spectral feature parameter extraction model, is the illumination intensity of the collection site, is the illumination intensity of the infrared spectrum image collection.

[0030] The formula for calculating the integral area environmental correction value is:

[0031]

[0032] In the formula, is the integral area environmental correction value, is the integral area prediction value of the to-be-detected object output by the spectral feature parameter extraction model.

[0033] Further, the environmental correction feature parameter is corrected based on the obtained physical characteristic parameter to obtain the spectral feature parameter accurate value of the to-be-detected object, wherein the formula for calculating the peak position accurate value is:

[0034]

[0035] In the formula, is the peak position accurate value, is the particle size distribution correction index, is the surface roughness correction index, and respectively are the weight coefficients of the particle size distribution correction index and the surface roughness correction index, wherein and and are both greater than 0;

[0036] wherein the formula for calculating the peak intensity accurate value is:

[0037]

[0038] In the formula, is the peak intensity accurate value, is the thickness correction index, is the weight coefficient of the sum of the particle size distribution correction index and the surface roughness correction index, is the weight coefficient of the thickness correction index, wherein and and are both greater than 0;

[0039] wherein the formula for calculating the integral area accurate value is:

[0040]

[0041] In the formula, is the integral area accurate value, is the weight coefficient of the sum of the particle size distribution correction index and the surface roughness correction index, is the weight coefficient of the thickness correction index, wherein and and are both greater than 0.

[0042] Further, wherein the specific formula for calculating the particle size distribution correction index, the surface roughness correction index and the thickness correction index is:

[0043]

[0044]

[0045]

[0046] In the formula, is the average diameter of the particle diameter of the object to be detected, is the wavelength of the incident infrared light, is the surface roughness of the object to be detected, is the thickness of the object to be detected;

[0047] wherein the surface roughness of the object to be detected The formula for calculation is:

[0048]

[0049]

[0050] In the formula, represents the sampling length of the stylus profilometer in the pth sampling area, p is the index of the sampling area, and , P is the number of sampling areas, is the height of the sampling point at x in the pth sampling area, x is the coordinate of the sampling point in the sampling length, is the surface roughness of the pth sampling area, is the surface roughness of the object to be detected, the center line and Data processing is carried out by the built-in software of the stylus profilometer.

[0051] The application also provides an infrared spectrum chemometrics analysis system for executing the above-mentioned infrared spectrum chemometrics analysis method, comprising:

[0052] A sample image processing module is configured to acquire infrared spectrum images of a plurality of known spectrum characteristic parameters, pre-process the collected infrared spectrum images to obtain a training sample image set, wherein the pre-processing includes baseline correction pre-processing and noise removal pre-processing, and the spectrum characteristic parameters include peak position, peak intensity and integral area.

[0053] A feature extraction model training module is configured to establish a neural network prediction model based on the obtained training sample image set, take the infrared spectrum images in the training sample image set as the input of the neural network prediction model, and take the corresponding spectrum characteristic parameters as labels, train the neural network prediction model, and obtain a spectrum characteristic parameter extraction model.

[0054] An environment correction module is configured to acquire an infrared spectrum image of the to-be-detected article, input the infrared spectrum image of the to-be-detected article into a spectrum feature parameter extraction model after pre-processing, output spectrum feature parameters of the to-be-detected article, and collect environment parameters when the infrared spectrum image of the to-be-detected article is acquired, wherein the environment parameters include an environment temperature of a collection place and an illumination intensity of the collection place.

[0055] An accurate feature value analysis module is configured to correct the output spectrum feature parameters of the to-be-detected article based on the collected environment parameters to obtain environment correction feature parameters, analyze physical characteristic parameters of the to-be-detected article, correct the environment correction feature parameters based on the obtained physical characteristic parameters to obtain accurate spectrum feature parameters of the to-be-detected article, complete chemical stoichiometric analysis of properties and contents of the to-be-detected article according to the obtained accurate spectrum feature parameters of the to-be-detected article, and wherein the physical characteristic parameters of the to-be-detected article include a mean particle size, a surface roughness and a thickness of the to-be-detected article.

[0056] Compared with the prior art, the present application has the following beneficial effects:

[0057] Firstly, the present application significantly reduces the spectrum distortion problem caused by spectrum baseline drift and environmental noise by pre-processing the collected infrared spectrum image, such as baseline correction and noise removal. This pre-processing process provides clearer and more accurate training data for subsequent neural network modeling, reduces the bias in the model training process, and improves the extraction accuracy of spectrum feature parameters including peak position, peak intensity and integral area. In addition, by establishing a neural network prediction model and training it based on a large number of training sample spectrum images and their known feature parameters, the complex nonlinear relationships and feature patterns in the spectrum data are automatically captured, greatly improving the accuracy and robustness of feature parameter extraction. Secondly, the present application innovatively includes the environment parameters including the environment temperature and the illumination intensity of the collection place in the correction process of the spectrum feature parameters. Finally, the environment corrected spectrum feature parameters are further corrected in combination with the physical characteristics of the to-be-detected sample, including the mean particle size, the surface roughness and the thickness. The influence of the physical characteristics on the spectrum feature parameters is fully considered, including the nonlinear modulation of the particle size distribution on the peak intensity and the peak position, the scattering effect of the surface roughness on the spectrum signal, and the complex influence of the thickness difference on the light absorption intensity. Through the correction process of the physical characteristic parameters, the present application can more accurately simulate the true value of the spectrum feature parameters, significantly improving the accuracy of the analysis results. BRIEF DESCRIPTION OF DRAWINGS

[0058] Figure 1 It is a whole method flowchart of the present application;

[0059] Figure 2 It is a whole system structure schematic diagram of the present application. DETAILED DESCRIPTION

[0060] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to specific embodiments.

[0061] It should be noted that, unless otherwise defined, technical terms or scientific terms used in the present application should be understood as their common meanings to those skilled in the art. The terms "first", "second", and similar terms used in the present application do not indicate any order, number, or importance, but are only used to distinguish different components. The terms "include", "contain", and similar terms mean that the elements or objects before the terms encompass the elements or objects listed after the terms and their equivalents, without excluding other elements or objects. The terms "connect" or "connected" and similar terms do not mean physical or mechanical connections, but can include electrical connections, whether direct or indirect. The terms "upper", "lower", "left", "right", and the like only indicate relative positional relationships, which can change accordingly when the absolute positions of the described objects change.

[0062] Embodiments:

[0063] Please refer to Figure 1 The present application provides a technical solution:

[0064] An infrared spectroscopy chemometrics analysis method, the specific steps comprising:

[0065] Step 1: Obtain a plurality of infrared spectrum images of known spectral characteristic parameters, and pre-process the collected infrared spectrum images to obtain a training sample image set, wherein the pre-processing includes baseline correction preprocessing and noise removal preprocessing, and the spectral characteristic parameters include peak position, peak intensity, and integral area.

[0066] Pre-process the collected infrared spectrum images to obtain a training sample image set, wherein the pre-processing includes baseline correction preprocessing and noise removal preprocessing, and a polynomial fitting method is selected for baseline correction.

[0067] The polynomial fitting method is a commonly used spectral baseline correction method, which is suitable for removing baseline drift in infrared spectrum data. In infrared spectrum analysis, due to instrument noise, environmental conditions, or physical properties of the sample itself, the baseline of the spectral signal may drift or change nonlinearly, resulting in reduced accuracy of extracting spectral characteristic parameters (such as peak position, peak intensity, etc.). Through the polynomial fitting method, the baseline can be effectively fitted and deducted from the spectral data, restoring the true spectral signal.

[0068] The original spectrum data requiring baseline correction is obtained, including spectral wavelength and corresponding spectral intensity. The spectral data is usually stored in the form of a two-dimensional array, wherein each data point is composed of wavelength and intensity value. The order of the fitting polynomial is selected; the baseline point is selected, which is the area in the spectrum that is not strongly affected by the signal characteristic peak, usually representing the true position of the baseline; at the selected baseline point, the baseline polynomial is fitted using the least squares method; the fitted baseline function is deducted from the original spectral intensity to obtain the corrected spectrum.

[0069] The specific method of the spectral characteristic parameter includes: using an algorithm (such as the first derivative method) to find the local maximum value of the spectral curve, and the peak position is the wavelength or wave number corresponding to the maximum value point;

[0070] The intensity corresponding to the local maximum value point in the spectral curve is directly taken by using spectral analysis software or programming tools such as Origin, MATLAB or LabVIEW, which is the peak intensity;

[0071] Numerical integration is performed in a specific wavelength range or wave number range, and the commonly used methods include: trapezoidal method: the integral region is divided into small trapezoids, which are added one by one, and the integral range is usually the full width half maximum of the peak or a specific wavelength / wave number window, so as to obtain the integral area characteristic parameter.

[0072] The noise removal preprocessing is performed by using the ensemble empirical mode decomposition method, and the specific steps include: adding different amplitudes of noise to the original spectrum, performing empirical mode decomposition for several times, averaging the IMFs obtained by several times of decomposition, retaining the low-frequency IMF component, and filtering out the high-frequency component as noise; wherein the polynomial fitting method and the ensemble empirical mode decomposition method are both conventional technical means, and the detailed operation steps are not described here.

[0073] The generation method of the training sample image set is: one-to-one mapping the preprocessed infrared spectrum image and the spectral characteristic parameter corresponding to the image to form a corresponding grid, and the formed grid is recorded as the training sample image set.

[0074] Step 2: Based on the obtained training sample image set, a neural network prediction model is established, the infrared spectrum image in the training sample image set is taken as the input of the neural network prediction model, and the corresponding spectral characteristic parameter is taken as the label, the neural network prediction model is trained, and the spectral characteristic parameter extraction model is obtained.

[0075] Based on the obtained training sample image set, a neural network prediction model is established, wherein a long short-term memory network model LSTM model is selected as the base model, an activation function and an optimization algorithm are selected, wherein a Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is:

[0076]

[0077] wherein, denotes the Tanh function, the argument denotes the input weighted sum of neurons, i.e. the result of the weighted sum of the inputs received by the neuron from the previous layer;

[0078] Meanwhile, the hyperparameters of the LSTM model are set, including the number of network layers, the number of iterations, the learning rate, the batch size, the number of training times, the batch quantity and the number of hidden layer neurons.

[0079] The number of network layers is set to a 3-layer network structure, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch quantity is set to 256, and the number of hidden layer neurons is set to 32.

[0080] The input of the trained spectral feature parameter extraction model is the preprocessed infrared spectrum image, and the output is the spectral feature parameter in the infrared spectrum image, including the peak position, the peak intensity and the integral area.

[0081] The infrared spectrum data often contains complex characteristic peaks and baseline drifts and other nonlinear changes. LSTM can capture these complex features through its gating mechanism and accurately extract effective information in the spectrum. LSTM does not need to explicitly define artificial features such as feature peak position or width, but automatically learns the internal rules of spectral features through training, reducing the influence of human intervention. The LSTM model not only accurately extracts the core features of the spectrum, but also has a certain robustness to noise or outliers through its dynamic memory mechanism, significantly improving the prediction accuracy of spectral feature parameters. Through the introduction of the LSTM model, the intelligentization and automation of spectral analysis can be realized, and the whole process from data input to result output can be completed without human intervention, which is suitable for efficient industrial production or rapid analysis in the laboratory.

[0082] Step 3: Obtain the infrared spectrum image of the to-be-detected object, pre-process the infrared spectrum image of the to-be-detected object, input it into the trained spectral feature parameter extraction model, output the spectral feature parameters of the to-be-detected object, and collect the environmental parameters when obtaining the infrared spectrum image of the to-be-detected object, including the environmental temperature of the collection place and the illumination intensity of the collection place.

[0083] Based on the collected environmental parameters, the output spectral feature parameters of the to-be-detected object are corrected to obtain environmental correction feature parameters, wherein the peak position is corrected by the environmental temperature to obtain a peak position environmental correction value, and the formula for calculating the peak position environmental correction value is:

[0084]

[0085] In the formula, is the peak position environmental correction value of the infrared spectrum image of the to-be-detected article, is the peak position prediction value of the to-be-detected article output by the spectral feature parameter extraction model, is the environmental temperature of the collection site, is the reference temperature of the infrared spectrum image collection.

[0086] The peak position of the infrared spectrum is closely related to the molecular vibration characteristics of the to-be-detected article and the environmental conditions. Changes in the environmental temperature will affect the position of the spectral peak through the following mechanism: when the temperature rises, the molecular thermal motion intensifies, which may cause the peak position to drift to a lower frequency (in the direction of lengthening the wavelength); when the temperature drops, the thermal motion weakens, and then the peak may move to a higher frequency (in the direction of shortening the wavelength), wherein reflects the proportional relationship between the current environmental temperature and the reference temperature. If , it indicates that the environmental temperature is higher than the reference temperature, otherwise it indicates that the environmental temperature is lower than the reference temperature.

[0087] It ensures the correctness of the correction direction: when the temperature rises, the peak may move to a lower frequency, and the correction needs to be increased; when the temperature drops, the peak may move to a higher frequency, and the correction needs to be reduced. indicates the relative amplitude of the temperature change and distinguishes the influence direction of the temperature rise or drop.

[0088] The peak intensity and the integral area are corrected by the environmental temperature and the illumination intensity of the collection site to obtain the peak intensity environmental correction value and the integral area environmental correction value, wherein the formula for calculating the peak intensity environmental correction value is:

[0089]

[0090] In the formula, is the peak intensity environmental correction value, is the peak intensity prediction value of the to-be-detected article output by the spectral feature parameter extraction model, is the illumination intensity of the collection site, is the illumination intensity of the infrared spectrum image collection.

[0091] The change of environmental temperature will change the activity level of molecular vibration and rotation. With the increase of temperature, the thermal motion of molecules is enhanced, which may lead to the decrease of the peak intensity of absorption spectrum; while the temperature decreases, the thermal motion is weakened, and the peak intensity may appear to be enhanced. The logarithmic function grows slowly, which is suitable for describing the nonlinear influence of temperature change on the peak intensity. That is, the greater the temperature change, the smaller the influence on the peak intensity, which conforms to the actual physical phenomenon. Since the increase of temperature usually leads to the decrease of peak intensity, the positive correction of the peak intensity with the increase of temperature is realized by .

[0092] The light intensity of the collection site will affect the signal-to-noise ratio of the detection instrument, and may cause the background noise of the spectrum to increase due to the superposition of scattered light or reflected light, thereby affecting the accuracy of the peak intensity. Therefore, the difference between the current collection light intensity and the reference light intensity is represented by , which directly reflects the influence of the change of light intensity on the peak intensity. The exponential function has the characteristics of rapid growth, which is suitable for describing the nonlinear influence of the change of light intensity on the peak intensity. That is, when the change of light intensity is small, the influence is relatively limited, while when the light intensity difference is large, the influence will be significantly amplified. Since the increase of light intensity usually causes the background noise to increase or the nonlinear drift of the instrument response to intensify, the negative correction of the peak intensity with the increase of light intensity is realized by .

[0093] The formula on which the integral area environmental correction value is calculated is:

[0094]

[0095] In the formula, is the integral area environmental correction value, is the integral area prediction value of the detected object output by the spectral feature parameter extraction model.

[0096] With the increase of temperature, the thermal vibration of molecules or atoms inside the material is enhanced, which may lead to changes in the peak intensity, position and width of some absorption peaks in the spectrum. In most cases, the increase of temperature will lead to the increase of the width of the absorption peak, because the thermal vibration of molecules or crystal lattices makes the distribution of different vibration energy levels or transition energy levels wider. In the absorption spectrum, the increase of temperature may lead to the generation of more excited states, thereby increasing the absorption intensity and thus increasing the integral area. Therefore, the negative correction of the integral area with the increase of temperature is realized by . In the formula, the term is subtracted, which means that the higher the temperature, the smaller the corrected integral area , making it more accurate to approach the actual feature.

[0097] Where the difference in light intensity has the same effect on the integral area as the peak intensity, an increase in light intensity will generally cause the integral area data to be biased too high, by subtracting a non-linear scaling of the square of the light intensity difference, thus correcting for the error introduced by the change in light intensity.

[0098] Step 4: Based on the collected environmental parameters, the output spectral characteristic parameters of the to-be-detected object are corrected to obtain environmental correction characteristic parameters, the physical characteristic parameters of the to-be-detected object are analyzed, the environmental correction characteristic parameters are corrected based on the obtained physical characteristic parameters to obtain accurate values of the spectral characteristic parameters of the to-be-detected object, and the chemical stoichiometric analysis of the properties and content of the to-be-detected object is completed according to the obtained accurate values of the spectral characteristic parameters of the to-be-detected object. The physical characteristic parameters of the to-be-detected object include the average particle size, surface roughness, and thickness of the to-be-detected object.

[0099] The environmental correction characteristic parameters are corrected based on the obtained physical characteristic parameters to obtain accurate values of the spectral characteristic parameters of the to-be-detected object, and the formula for calculating the accurate value of the peak position is:

[0100]

[0101] In the formula, is the accurate value of the peak position, is the particle size distribution correction index, is the surface roughness correction index, and are the weight coefficients of the particle size distribution correction index and the surface roughness correction index, respectively, where and are greater than 0.

[0102] When the particle size distribution becomes larger (the particle size increases), the scattering behavior changes from Rayleigh scattering (the particle size is much smaller than the wavelength) to Mie scattering (the particle size is close to the wavelength), that is, the increase in particle size will cause the peak position in the spectrum to redshift (move to a lower frequency), so the accurate value of the peak position should be positively compensated, and represents the positive compensation. Because the overall increases monotonically with the increase of , the amplitude of the accurate value of the peak position gradually increases.

[0103] ​​​When the surface roughness increases, the diffuse reflection component of light increases, and the specular reflection component decreases. This change has little effect on the peak position of the spectrum, but in some cases, it can cause a slight blue shift (moving to the high frequency direction) of the peak position, so the exact value of the peak position should be negatively compensated by , which represents the positive compensation. The square root function makes the correction term change faster when is larger, and change slower when is smaller.

[0104] where the particle size directly affects the scattering behavior of light, especially in light scattering and absorption dominated spectral measurements, the change of particle size will significantly change the scattering mode (such as Rayleigh scattering or Mie scattering), resulting in a significant shift of the spectral peak. With the increase of the particle size distribution index , the spectral peak position may change significantly, so a higher weight needs to be given, so and and are greater than 0.

[0105] where the formula for calculating the exact value of the peak intensity is:

[0106]

[0107] In the formula, is the exact value of the peak intensity, is the thickness correction index, is the weight coefficient of the sum of the particle size distribution correction index and the surface roughness correction index, is the weight coefficient of the thickness correction index, where and and are greater than 0;

[0108] The larger the particle size, the stronger the scattering, and more light is scattered and cannot enter the detector, resulting in a gradual decrease in peak intensity, so it needs to be positively corrected to make it closer to the true value.

[0109] The greater the surface roughness, the more significant the influence of the surface microstructure on the scattering and reflection of light, resulting in part of the light being scattered or reflected without entering the detector, thereby reducing the peak intensity. Therefore, the particle size distribution correction index and the surface roughness correction index have a synergistic effect on the peak intensity, so represents the correction of the particle size distribution and the surface roughness to the peak intensity. Their influence usually causes the peak intensity to decrease due to the scattering, reflection and absorption of light, so a positive compensation correction is needed.

[0110] The greater the thickness of the sample, the longer the propagation path of the light in the sample, the more significant the absorption effect, resulting in the weakening of the transmitted light intensity, thereby reducing the peak intensity of the spectrum. The influence of the sample thickness is usually significant, especially in spectral measurements, where the change in thickness will cause an exponential decay of the light intensity, so by the peak intensity is accurately determined. the forward correction is completed.

[0111] The change in thickness directly affects the length of the light propagation path, thereby significantly affecting the light absorption. Although the changes in particle size distribution and surface roughness will affect the light scattering and reflection, their effects are usually local, so the thickness is set as and and are both greater than 0.

[0112] wherein the formula for calculating the integral area accuracy value is:

[0113]

[0114] wherein, is the integral area accuracy value, is the weight coefficient of the sum of the particle size distribution correction index and the surface roughness correction index, is the weight coefficient of the thickness correction index, wherein and and are both greater than 0.

[0115] The greater the thickness of the sample, the longer the propagation path of the light in the sample, the stronger the light absorption, resulting in a decrease in the integral area of the transmitted light in the spectrum, so needs to be forward corrected. The influence of the thickness on the light transmission is usually nonlinear, which is manifested as the speed of absorption increasing gradually slows down with the increase of the thickness, so by taking the logarithm the nonlinearity is reflected.

[0116] The larger the particle size, the stronger the light scattering, and more light is scattered out of the sample or away from the detector, resulting in a decrease in the light intensity received by the detector, thereby reducing the integral area. The greater the surface roughness, the stronger the scattering and reflection effects, and part of the light is not measured, resulting in a decrease in the integral area, so the square root function reflects the nonlinear forward compensation effect of light scattering and surface roughness on the integral area.

[0117] The change in thickness directly affects the length of the light propagation path, thereby significantly changing the light absorption intensity. The changes in particle size distribution and surface roughness will only affect the light scattering and reflection, although they will also change the integral area, but their effects are usually secondary, so the thickness is set as and and All are greater than 0.

[0118] The specific formula for calculating the particle size distribution correction index, the surface roughness correction index and the thickness correction index is as follows:

[0119]

[0120]

[0121]

[0122] In the formula, is the average diameter of the particle size of the to-be-detected article, is the wavelength of the incident infrared light, is the surface roughness of the to-be-detected article, is the thickness of the to-be-detected article.

[0123] The surface roughness of the to-be-detected article is represented by the average value of the surface roughness of the plurality of sampling regions, and the specific surface roughness of the to-be-detected article is The formula for calculating is as follows:

[0124]

[0125]

[0126] In the formula, represents the sampling length of the stylus profilometer in the pth sampling region, p is the index of the sampling region, and P is the number of sampling regions, is the height of the sampling point at x in the sampling length of the pth sampling region, x is the coordinate of the sampling point in the sampling length, is the surface roughness of the pth sampling region, is the surface roughness of the to-be-detected article, the center line and is obtained by data processing by the built-in software of the stylus profilometer.

[0127] Please refer to Figure 2 The application also provides an infrared spectrum chemometrics analysis system for executing the above-mentioned infrared spectrum chemometrics analysis method, comprising:

[0128] The sample image processing module is configured to acquire infrared spectrum images of a plurality of known spectrum characteristic parameters, pre-process the collected infrared spectrum images to obtain a training sample image set, wherein the pre-processing includes baseline correction pre-processing and noise removal pre-processing, and the spectrum characteristic parameters include peak position, peak intensity and integral area.

[0129] The feature extraction model training module is configured to establish a neural network prediction model based on the obtained training sample image set, take the infrared spectrum image in the training sample image set as the input of the neural network prediction model, and take the corresponding spectrum feature parameter as the label to train the neural network prediction model, and obtain the spectrum feature parameter extraction model.

[0130] The environment correction module is configured to obtain the infrared spectrum image of the to-be-detected object, input the preprocessed infrared spectrum image of the to-be-detected object into the trained spectrum feature parameter extraction model, output the spectrum feature parameter of the to-be-detected object, and collect the environment parameter when the infrared spectrum image of the to-be-detected object is obtained, wherein the environment parameter includes the environment temperature of the collection place and the illumination intensity of the collection place.

[0131] The accurate feature value analysis module is configured to correct the output spectrum feature parameter of the to-be-detected object based on the collected environment parameter to obtain the environment correction feature parameter, analyze the physical characteristic parameter of the to-be-detected object, correct the environment correction feature parameter based on the obtained physical characteristic parameter to obtain the accurate spectrum feature parameter of the to-be-detected object, and complete the chemical stoichiometric analysis of the property and content of the to-be-detected object according to the obtained accurate spectrum feature parameter of the to-be-detected object, wherein the physical characteristic parameter of the to-be-detected object includes the average particle size, the surface roughness and the thickness of the to-be-detected object.

[0132] The above formulas are all dimensionless numerical calculations, the formulas are obtained by software simulation of a large amount of data to obtain a formula of the nearest real situation, and the preset parameters in the formulas are set by a person skilled in the art according to the actual situation.

[0133] The above embodiments can be realized wholly or partially by software, hardware, firmware or any combination thereof. When realized by software, the above embodiments can be realized in the form of a computer program product wholly or partially. Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized by hardware or software methods depends on the specific application and design constraints of the technical solutions.

[0134] The units described as separate components can or can not be physically separate, and the components shown as units can or can not be physical units, which can be located in one place or distributed on multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiments.

[0135] The above merely provides the specific implementation of the present application, but the protection scope of the present application is not limited to this. Any person skilled in the art can easily think of the changes or replacements within the technical range disclosed by the present application, which should be covered in the protection scope of the present application.

Claims

1. An infrared spectroscopic chemometric analysis method, characterized in that: The specific steps include: Acquire a plurality of infrared spectral images with known spectral characteristic parameters, and preprocess the acquired infrared spectral images to obtain a training sample image set, wherein the preprocessing includes baseline correction preprocessing and noise removal preprocessing, and the spectral characteristic parameters include peak position, peak intensity, and integrated area; Based on the obtained training sample image set, a neural network prediction model is established. The infrared spectrum images in the training sample image set are used as inputs of the neural network prediction model. The corresponding spectral feature parameters are used as labels to train the neural network prediction model and obtain a spectral feature parameter extraction model. Acquire an infrared spectral image of the object to be detected, preprocess the infrared spectral image of the object to be detected, input it into a trained spectral feature parameter extraction model, and output the spectral feature parameters of the object to be detected. At the same time, collect environmental parameters when acquiring the infrared spectral image of the object to be detected, the environmental parameters including the ambient temperature and light intensity of the collection location; Based on the collected environmental parameters, the output spectral characteristic parameters of the object to be detected are corrected to obtain environmentally corrected characteristic parameters, wherein the peak position is specifically corrected by the ambient temperature to obtain the peak position environmental correction value, and the peak intensity and integrated area are corrected by the ambient temperature and the light intensity of the collection location to obtain the peak intensity environmental correction value and the integrated area environmental correction value; the physical property parameters of the object to be detected are analyzed, and the environmental correction characteristic parameters are corrected based on the obtained physical property parameters to obtain the precise values ​​of the spectral characteristic parameters of the object to be detected. According to the obtained precise values ​​of the spectral characteristic parameters of the object to be detected, the chemometric analysis of the properties and content of the object to be detected is completed, and the physical property parameters of the object to be detected include the mean particle size, surface roughness and thickness of the object to be detected.

2. The infrared spectroscopic chemometric analysis method according to claim 1, wherein: The collected infrared spectrum images are preprocessed to obtain a training sample image set, wherein the preprocessing includes baseline correction preprocessing and noise removal preprocessing, wherein a polynomial fitting method is selected for baseline correction, and the specific operation steps include: fitting a polynomial by a least squares method to obtain baseline parameters, and then subtracting the baseline signal from the original spectrum data to obtain a corrected spectrum; The noise removal preprocessing is performed using the ensemble empirical mode decomposition method. The specific steps include: adding noise of different amplitudes to the original spectrum, performing several empirical mode decompositions, averaging the IMF components obtained from the several decompositions, retaining the low-frequency IMF components, and filtering out the high-frequency components as noise; The method for generating the training sample image set is as follows: mapping the preprocessed infrared spectrum image to the spectral characteristic parameters corresponding to the image one by one to form a corresponding grid, and recording the formed grid as the training sample image set.

3. An infrared spectroscopic chemometric analysis method according to claim 2, characterized in that: Based on the obtained training sample image set, a neural network prediction model is established, in which the long short-term memory network model LSTM model is selected as the base model, and the activation function and optimization algorithm are selected. The Tanh function is selected as the activation function, and Adam is selected as the optimization algorithm of the LSTM model; the formula of the Tanh function is: ； Where, Represents the Tanh function, independent variable Represents the weighted sum of the neuron's input, that is, the result of weighted summation of the input received by the neuron from the previous layer; At the same time, the hyperparameters of the LSTM model are set, including the number of network layers, number of iterations, learning rate, batch size, number of training times, batch size, and number of hidden layer neurons; The number of network layers is set to 3 layers, the number of iterations is set to 200, the learning rate is set to 0.001, the batch size is set to 32, the number of training times is set to 100, the batch size is set to 256, and the number of hidden layer neurons is 32; The input of the trained spectral feature parameter extraction model is the preprocessed infrared spectral image, and the output is the spectral feature parameters in the infrared spectral image, including peak position, peak intensity and integrated area.

4. The infrared spectroscopic chemometric analysis method according to claim 3, wherein: The formula for calculating the peak position environmental correction value is: ； Where, is the peak position environmental correction value of the infrared spectrum image of the object to be detected, is the peak position prediction value of the object to be detected output by the spectral feature parameter extraction model, is the ambient temperature of the collection site, The reference temperature for infrared spectrum image acquisition.

5. An infrared spectroscopic chemometric analysis method according to claim 4, characterized in that: The peak intensity and integrated area are corrected by the ambient temperature and the light intensity of the collection site to obtain the peak intensity environmental correction value and the integrated area environmental correction value. The formula for calculating the peak intensity environmental correction value is: ； Where, is the peak intensity environmental correction value, is the peak intensity prediction value of the object to be detected output by the spectral feature parameter extraction model, is the light intensity at the collection site, Light intensity collected for infrared spectral images; The formula for calculating the integrated area environmental correction value is: ； Where, is the environmental correction value of the integrated area, It is the predicted value of the integrated area of ​​the object to be detected output by the spectral feature parameter extraction model.

6. The infrared spectroscopic chemometric analysis method according to claim 5, wherein: Based on the obtained physical characteristic parameters, the environmental correction characteristic parameters are corrected to obtain the precise value of the spectral characteristic parameters of the object to be detected. The formula for calculating the precise value of the peak position is: ； Where, is the exact value of the peak position, is the particle size distribution correction index, is the surface roughness correction index, and are the weight coefficients of the particle size distribution correction index and the surface roughness correction index, respectively, where and and All greater than 0; The formula for calculating the exact value of peak intensity is: ； Where, is the exact value of the peak intensity, is the thickness correction index, is the weight coefficient of the sum of the particle size distribution correction index and the surface roughness correction index, is the weight coefficient of the thickness correction index, where and and All greater than 0; The formula for calculating the exact value of the integral area is: ； Where, is the exact value of the integrated area, is the weight coefficient of the sum of the particle size distribution correction index and the surface roughness correction index, is the weight coefficient of the thickness correction index, where and and Both are greater than 0.

7. The infrared spectroscopic chemometric analysis method according to claim 6, wherein: The specific formulas for calculating the particle size distribution correction index, surface roughness correction index, and thickness correction index are as follows: ； ； ； Where, is the average diameter of the particles to be tested, is the wavelength of the incident infrared light, is the surface roughness of the object to be tested, is the thickness of the object to be tested; Among them, the surface roughness of the object to be tested The calculation is based on the formula: ； ； Where, represents the sampling length of the stylus profilometer in the pth sampling area, where p is the index of the sampling area, and , P is the number of sampling areas, is the height of the xth sampling point within the sampling length of the pth sampling area from the center line, x is the coordinate of the sampling point within the sampling length, is the surface roughness of the pth sampling area, is the surface roughness, centerline and The data is obtained by processing the built-in software of the stylus profilometer.

8. An infrared spectroscopic chemometric analysis system, characterized in that: The infrared spectroscopic chemometric analysis system is used to perform the infrared spectroscopic chemometric analysis method according to any one of claims 1 to 7, comprising: a sample image processing module, configured to acquire a plurality of infrared spectral images with known spectral characteristic parameters, and preprocess the acquired infrared spectral images to obtain a training sample image set, wherein the preprocessing includes baseline correction preprocessing and noise removal preprocessing, and the spectral characteristic parameters include peak position, peak intensity, and integrated area; The feature extraction model training module is used to establish a neural network prediction model based on the obtained training sample image set, use the infrared spectrum images in the training sample image set as the input of the neural network prediction model, and use the corresponding spectral feature parameters as labels to train the neural network prediction model to obtain a spectral feature parameter extraction model; An environmental correction module is used to obtain an infrared spectral image of the object to be detected, pre-process the infrared spectral image of the object to be detected, input the image into a trained spectral feature parameter extraction model, output the spectral feature parameters of the object to be detected, and simultaneously collect environmental parameters when the infrared spectral image of the object to be detected is obtained, including the ambient temperature and light intensity of the collection location; The precise characteristic value analysis module is used to correct the output spectral characteristic parameters of the object to be detected based on the collected environmental parameters to obtain environmentally corrected characteristic parameters, wherein the peak position is corrected by the ambient temperature to obtain the peak position environmental correction value, and the peak intensity and integrated area are corrected by the ambient temperature and the light intensity of the collection location to obtain the peak intensity environmental correction value and the integrated area environmental correction value; the physical property parameters of the object to be detected are analyzed, and the environmental correction characteristic parameters are corrected based on the obtained physical property parameters to obtain the precise values ​​of the spectral characteristic parameters of the object to be detected; based on the obtained precise values ​​of the spectral characteristic parameters of the object to be detected, a chemometric analysis of the properties and content of the object to be detected is completed, wherein the physical property parameters of the object to be detected include the mean particle size, surface roughness and thickness of the object to be detected.

Images