Method and system for detecting average atomic number of material sample, medium, equipment and product

Through backscattered electron image preprocessing, K-mean clustering model segmentation and nonlinear least squares fitting, the average atomic number of materials is directly obtained, solving the problems of low detection accuracy and complex operation in the prior art, and achieving efficient and convenient detection of non-magnetic elements, compounds and mixtures.

CN120411960APending Publication Date: 2025-08-01NANJING NORMAL UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202510373227.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-03-27
Publication Date
2025-08-01

AI Technical Summary

Technical Problem

The existing methods for detecting average atomic number of material samples have problems such as low detection accuracy, complex operation and high cost, especially the analysis and detection effect of non-magnetic elements, compounds and mixtures.

Method used

The methods of backscattered electron image preprocessing, K-mean clustering model segmentation and nonlinear least squares fitting are used to directly obtain the average atomic number of materials through image preprocessing, image segmentation and data fitting, which simplifies the detection process and improves the accuracy.

Benefits of technology

It realizes efficient and convenient detection of non-magnetic elements, compounds and mixtures, simplifies the operation process, improves detection accuracy and efficiency, and reduces costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120411960A_ABST
    Figure CN120411960A_ABST
Patent Text Reader

Abstract

The invention discloses an average atomic number detection method and system for a material sample, a medium, equipment and a product in the technical field of scanning electron microscope detection, and aims to solve the optimization problem of the existing detection technology. The method comprises the following steps: carrying out image preprocessing on an obtained backscattered electron image of a to-be-detected material sample to obtain a processed image; performing image segmentation on the processed image by using a trained clustering model to obtain a to-be-tested material sample area; calculating the average gray value of the to-be-detected material sample area of the processed image; and inputting the average gray value of the to-be-tested material sample area of the processed image into the fitted index model to obtain the predicted average atomic number of the to-be-tested material sample. According to the invention, characterization of material elements and distribution states thereof can be realized efficiently and conveniently, and the method is suitable for analysis and detection of various non-magnetic elementary substances, compounds and mixtures.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of scanning electron microscope detection, and particularly to a method, system, medium, device and product for detecting the average atomic number of a material sample. Background Art

[0002] A scanning electron microscope is a widely used analytical and testing tool that obtains multi-dimensional information of a sample by scanning the surface of the sample with an electron beam and detecting the signals generated by the interaction with the sample. These signals mainly include secondary electrons, backscattered electrons, Auger electrons, characteristic X-rays, etc. Each type of signal provides different characteristics of the sample. For example, secondary electrons are used to observe the microscopic morphology of the sample surface; characteristic X-rays can qualitatively and quantitatively analyze the elemental composition of the sample through the correlation between energy and atomic species; transmission electrons are used to study the crystal structure, lattice constant, and crystal defects of materials.

[0003] Among these physical signals, backscattered electrons have the characteristic of being closely related to the average atomic number of the sample. Backscattered electrons originate from the region 10 nm to 1 μm below the sample surface, and their generation is affected by both the incident electron energy and the average atomic number (Z) of the sample. Since the atomic nuclei of heavy elements have a stronger scattering ability for electrons, under the condition that the acceleration voltage and current are kept constant, the larger the average atomic number of the sample, the more backscattered electrons are generated. In the backscattered electron image, when the brightness and contrast are kept constant, the larger the average atomic number of the sample, the brighter the image; in the backscattered electron image after grayscale conversion and normalization, the grayscale value increases with the increase of the atomic number. Therefore, by analyzing the grayscale value distribution of the normalized backscattered electron image, the average atomic number and its distribution state of the sample can be intuitively and quantitatively reflected.

[0004] Existing detection methods are diverse, but each has its limitations. For example, Raman spectroscopy can perform non-destructive analysis, but the detection results are easily interfered by the fluorescence background, and the detection accuracy depends on the optical properties of the material; electron energy loss spectroscopy has high detection accuracy, but requires ultra-thin slices of the sample, with complex operations and high costs. Summary of the Invention

[0005] The purpose of the present invention is to overcome the deficiencies in the prior art and provide a method, system, medium, device and product for detecting the average atomic number of a material sample, which can efficiently and conveniently realize the characterization of material elements and is applicable to the analysis and detection of various non-magnetic simple substances, compounds and mixtures.

[0006] To solve the above technical problems, the present invention is implemented by adopting the following technical solutions:

[0007] In the first aspect, the present invention provides a method for detecting the average atomic number of a material sample, including:

[0008] Perform image preprocessing on the backscattered electron image of the material sample to be measured, and obtain the processed image;

[0009] Use the trained clustering model to perform image segmentation on the processed image to obtain the area of the material sample to be measured;

[0010] Calculate the average gray value of the area of the material sample to be measured in the processed image;

[0011] Input the average gray value of the area of the material sample to be measured in the processed image into the fitted exponential model to obtain the predicted average atomic number of the material sample to be measured.

[0012] Optionally, the image preprocessing includes gray conversion, normalization processing, and Gaussian smoothing processing.

[0013] Optionally, the clustering model adopts the K-means clustering algorithm, and the training process of the clustering model includes:

[0014] Obtain the training set of backscattered electron images of the sample material;

[0015] Perform image preprocessing on the training set of backscattered electron images of the sample material to obtain the processed sample set;

[0016] For each sample in the processed sample set, repeat the following training operations until the preset stop condition is met:

[0017] Optionally select two pixel points in the sample as the initial clustering centers of the sample area class and the background area class respectively;

[0018] Calculate the distances from all pixel points in the sample to the clustering centers of the sample area class and the background area class respectively, and assign each pixel point to the class to which the clustering center with the closest distance belongs;

[0019] Take the mean value of all pixel points in the sample area class as the new clustering center of the sample area, and take the mean value of all pixel points in the background area as the new clustering center of the background area class;

[0020] After all samples in the processed sample set are trained, obtain the trained clustering model.

[0021] Optionally, using the trained clustering model to perform image segmentation on the processed image includes:

[0022] Calculate the distances from each pixel point in the processed image to the clustering centers of the sample area class and the background area class respectively, and assign each pixel point to the class to which the clustering center with the closest distance belongs;

[0023] Perform image segmentation on the processed image according to all pixel points in the sample area class to obtain the material sample area.

[0024] Optionally, the exponential model uses the nonlinear least squares method for data fitting, and the data fitting process of the exponential model includes:

[0025] Obtain a data fitting training set;

[0026] According to the data fitting training set, establish an average atomic number set and an average gray value set, and store the average atomic number of each sample in the data fitting training set and the average gray value of the material sample area of each sample respectively;

[0027] Initialize the fitting parameters of the exponential model;

[0028] Repeat the following operations until the sum of squared residuals of the exponential model is less than the threshold to obtain a well-fitted exponential model:

[0029] According to the average atomic number set and the average gray value set, use a nonlinear optimization algorithm to solve the increment of the fitting parameters of the exponential model, and update the fitting parameters according to the increment of the fitting parameters;

[0030] According to the average atomic number set and the average gray value set, calculate the sum of squared residuals of the exponential model.

[0031] Optionally, the average atomic number of each sample in the data fitting training set is calculated by the following formula:

[0032] where represents the average atomic number of the sample, represents the number of elements in the sample, represents the th element in the sample, represents the th element in the sample;

[0033] The calculation formula of the exponential model is as follows:

[0034]

[0035] where represents the predicted average atomic number output by the exponential model, represents the average gray value of the material sample area, represents the fitting parameters of the exponential model;

[0036] The sum of squared residuals The calculation formula is as follows:

[0037]

[0038] Wherein, represents the number of samples in the data fitting training set, represents the th average atomic number of the samples in the data fitting training set, represents the average gray value of the material sample area of the th sample in the data fitting training set;

[0039] The update of the fitting parameters is achieved through the following formula:

[0040]

[0041] Wherein, represents the number of updates of the exponential model, represents the fitting parameter of the th update of the exponential model, represents the fitting parameter of the th update of the exponential model, represents the increment of the fitting parameter of the exponential model.

[0042] In a second aspect, the present invention provides an average atomic number detection system for a material sample, comprising:

[0043] An image preprocessing module, configured to: perform image preprocessing on the backscattered electron image of the material sample to be measured obtained, to obtain a processed image;

[0044] An image segmentation module, configured to: use the trained clustering model to perform image segmentation on the processed image, to obtain the area of the material sample to be measured;

[0045] An average gray value calculation module, configured to: calculate the average gray value of the area of the material sample to be measured in the processed image;

[0046] A data fitting module, configured to: input the average gray value of the area of the material sample to be measured in the processed image into the fitted exponential model, to obtain the predicted average atomic number of the material sample to be measured.

[0047] In a third aspect, the present invention provides a computer-readable storage medium, on which computer instructions are stored, and when the computer instructions are executed by a processor, the steps of the average atomic number detection method for a material sample according to any one of the first aspects are implemented.

[0048] In a fourth aspect, the present invention provides a computer device, comprising:

[0049] A memory for storing computer instructions;

[0050] A processor for executing the computer instructions to implement the steps of the method for detecting the average atomic number of a material sample according to any one of the first aspect.

[0051] In a fifth aspect, the present invention provides a computer program product comprising computer instructions, characterized in that when the computer instructions are executed by a processor, the steps of the method for detecting the average atomic number of a material sample according to any one of the first aspect are implemented.

[0052] Compared with the prior art, the beneficial effects achieved by the present invention:

[0053] 1. The method for detecting the average atomic number of a material sample provided by the present invention is different from the secondary electron technology of traditional microscopes and scanning electron microscopes. This method can directly obtain the average atomic number of the material. Compared with energy spectrum analysis technology, this method judges the characteristics of the material by quantifying the average gray value of the backscattered electron image, which is more simple and efficient and has significant practical value.

[0054] 2. The system for detecting the average atomic number of a material sample provided by the present invention can efficiently and conveniently realize the characterization of material elements by setting an image preprocessing module, an image segmentation module, a data fitting module and a material characterization module. It is applicable to the analysis and detection of various non-magnetic simple substances, compounds and mixtures, and has practical significance and good application prospects.

[0055] 3. The computer-readable storage medium, computer device and computer program product provided by the present invention can execute the steps of the method for detecting the average atomic number of a material sample provided by the present invention. Description of the Drawings

[0056] Figure 1 It is a flowchart of the method for detecting the average atomic number of a material sample according to an embodiment of the present invention;

[0057] Figure 2 It is a secondary electron image and a backscattered electron image according to an embodiment of the present invention;

[0058] Figure 3 It is a diagram of the image segmentation result of the clustering model based on the K-means clustering algorithm according to an embodiment of the present invention;

[0059] Figure 4 It is a relationship curve diagram of the average gray value of the experimental sample and the average atomic number predicted by the exponential model according to an embodiment of the present invention. Detailed Embodiments

[0060] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific features in the embodiments of the present application and the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. Without conflict, the technical features in the embodiments of the present application and the embodiments can be combined with each other.

[0061] It should be noted that the term "and / or" in this article is merely a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this article generally represents an "or" relationship between the associated objects before and after.

[0062] Embodiment 1:

[0063] The embodiment of the present invention discloses a method for detecting the average atomic number of a material sample. Referring to Figure 1 as shown, the specific steps are as follows:

[0064] S1. Perform image preprocessing on the backscattered electron image of the material sample to be measured obtained, and obtain the processed image;

[0065] S2. Use the trained clustering model to perform image segmentation on the processed image to obtain the area of the material sample to be measured;

[0066] S3. Calculate the average gray value of the area of the material sample to be measured in the processed image;

[0067] S4. Input the average gray value of the area of the material sample to be measured in the processed image into the fitted exponential model to obtain the predicted average atomic number of the material sample to be measured.

[0068] Specifically, in step S1, the backscattered electron image is obtained through the following steps (in this embodiment, the sample materials for training include BN, MoS2, Cr, etc., and the backscattered electron images of the sample materials are also obtained through the following steps):

[0069] (1) Test preparation

[0070] Transfer the sample to the substrate by pressing to ensure that the sample is in close contact with the substrate. Gently blow the surface of the sample with an ear bulb to ensure that the surface is clean and flat; place the sample in the scanning electron microscope cavity and start the vacuum pumping system; after about 10 minutes, turn on the electron beam and set the acceleration voltage to 5 keV and the current to 30 pA;

[0071] (2) Image acquisition

[0072] First, use a secondary electron detector to adjust parameters such as brightness, contrast, and astigmatism in the secondary electron mode, and focus on the sample until a clear secondary electron image is obtained;

[0073] Next, insert a backscattered electron detector, adjust parameters such as brightness, contrast, and astigmatism in the backscattered electron mode, ensure that all setting parameters are stable, and focus on the sample again until a clear backscattered electron image is obtained.

[0074] In this embodiment, the substrate used for the sample is a non-magnetic material, which is used to support the sample and is required to have a smooth, clean, and pollution-free surface; in this embodiment, a wafer with silicon dioxide on the upper layer and silicon on the lower layer is selected as the substrate; the acceleration voltage of the scanning electron microscope is preferably set to 5 keV to meet the imaging requirements under experimental conditions; when collecting secondary electron images, according to the characteristic requirements of different samples, the brightness and contrast parameters can be adjusted to obtain the optimal image effect; when collecting backscattered electron images, unified parameter settings should be used for different samples to keep the brightness and contrast parameters constant to ensure the reliability and comparability of the data; when collecting secondary electron images and backscattered electron images, the working distance and magnification should be kept consistent, while the brightness and contrast can be appropriately adjusted; in this embodiment, the brightness and contrast parameters of the backscattered electron mode are set to 75% and 68.5% respectively.

[0075] The image preprocessing includes grayscale conversion, normalization processing, and Gaussian smoothing processing; in the backscattered electron image, elements with different atomic numbers have different absorption rates for the electron beam energy, thus showing different light and dark differences; by performing normalization processing on the grayscale backscattered electron image, elements with different atomic numbers will show different grayscale values; through the data distribution of the grayscale values, the average atomic number and its distribution state of the sample can be deduced; therefore, in this embodiment, the image preprocessing includes using Matlab software to convert the backscattered electron image into a grayscale image, and performing normalization processing and Gaussian smoothing processing on the grayscale image, and the standard deviation of the Gaussian smoothing processing is set to 2.

[0076] In step S2, the clustering model uses the K-means clustering algorithm to perform image segmentation on the preprocessed image, and divides the pixel points of the processed image into two categories, which are respectively labeled as the sample material label and the background label; refer to Figure 3 The figure shows a binary image created according to the sample material label and the background label. The white part is the area of the sample material to be measured, and the black part is the background area; the training process of the clustering model includes:

[0077] Obtain a training set of backscattered electron images of the sample material;

[0078] Perform image preprocessing on the training set of backscattered electron images of the sample material to obtain a processed sample set;

[0079] For each sample in the processed sample set, the following training operations are repeatedly executed until a preset stop condition is satisfied:

[0080] Arbitrarily select two pixel points in the sample as the initial clustering centers of the sample region class and the background region class respectively;

[0081] Calculate the distances from all pixel points in the sample to the clustering center of the sample region class and the clustering center of the background region class respectively, and assign each pixel point to the class to which the clustering center with the closest distance belongs;

[0082] Take the mean value of all pixel points in the sample region class as the new clustering center of the sample region, and take the mean value of all pixel points in the background region as the new clustering center of the background region class;

[0083] After all samples in the processed sample set are trained, a trained clustering model is obtained.

[0084] In this embodiment, using the trained clustering model, the image segmentation of the processed image includes:

[0085] Calculate the distances from each pixel point in the processed image to the clustering center of the sample region class and the clustering center of the background region class respectively, and assign each pixel point to the class to which the clustering center with the closest distance belongs;

[0086] According to all pixel points in the sample region class, perform image segmentation on the processed image to obtain the material sample region.

[0087] In step S3, the exponential model uses the nonlinear least squares method for data fitting, and the fitting objective is to find appropriate fitting parameters so that the sum of squared residuals of the exponential model is less than 0.05; the data fitting process of the exponential model includes:

[0088] Obtain the data fitting training set;

[0089] According to the data fitting training set, establish an average atomic number set and an average gray value set, and store the average atomic number of each sample and the average gray value of the material sample region of each sample in the data fitting training set respectively;

[0090] Initialize the fitting parameters of the exponential model (in this embodiment, set a = 1, b = 1, c = 1);

[0091] Repeat the following operations until the sum of squared residuals of the exponential model is less than the threshold (the threshold is 0.05) to obtain a well-fitted exponential model:

[0092] According to the average atomic number set and the average gray value set, use a non-linear optimization algorithm (such as the Levenberg-Marquardt method or the Gauss-Newton method) to solve for the increment of the fitting parameters of the exponential model, and update the fitting parameters according to the increment of the fitting parameters;

[0093] Calculate the sum of squared residuals of the exponential model according to the average atomic number set and the average gray value set.

[0094] The average atomic number of each sample in the data fitting training set is calculated by the following formula:

[0095]

[0096] where represents the average atomic number of the sample, represents the number of elements in the sample, represents the th element in the sample, represents the th element in the sample,

[0097] The calculation formula of the exponential model is as follows:

[0098]

[0099] where represents the predicted average atomic number output by the exponential model, represents the average gray value of the material sample area, represents the fitting parameters of the exponential model;

[0100] The sum of squared residuals of the exponential model is calculated by the following formula:

[0101]

[0102] where represents the number of samples in the data fitting training set, represents the th sample in the data fitting training set, represents the th sample in the data fitting training set,

[0103] The update of the fitting parameters is achieved by the following formula:

[0104]

[0105] where represents the update times of the exponential model, Denote the fitting parameters of the th update of the exponential model, Denote the fitting parameters of the th update of the exponential model, Denote the increment of the fitting parameters of the exponential model.

[0106] In this embodiment, after the model converges, the updated fitting parameters are output, and the following exponential model formula is obtained:

[0107] .

[0108] To test the effectiveness and accuracy of the exponential model, in this embodiment, the fitted exponential model is used to test elemental sulfur (S); the sum of squared fitting residuals of the exponential model obtained through experiments is 0.0209, which is less than 0.05, indicating that the initial parameter setting is reasonable and the fitting effect of the exponential model and the experimental data is good; the average gray value of the sample material region of elemental sulfur (S) obtained through image preprocessing and image segmentation is 0.3941, the true average atomic number of elemental sulfur (S) is 16, and the predicted average atomic number obtained by inputting the average gray value of 0.3941 into the exponential model is 14.43, and the prediction error is only 0.1%; this indicates that the exponential model can accurately predict the average atomic number of the sample, and the method for detecting the average atomic number of the material sample proposed in this example has a small error and high feasibility.

[0109] Figure 2 Shown are the secondary electron images and backscattered electron images of the obtained materials BN, MoS2, and Bi2O2Se; Figure 2 As shown in (a) is the secondary electron image of the material BN, Figure 2 As shown in (b) is the secondary electron image of the material MoS2, Figure 2 As shown in (c) is the secondary electron image of the material Bi2O2Se, Figure 2 As shown in (d) is the backscattered electron image of the material BN, Figure 2 As shown in (e) is the backscattered electron image of the material MoS2, Figure 2 As shown in (f) is the backscattered electron image of the material Bi2O2Se.

[0110] Figure 3 Shown are the results of image segmentation of the materials BN, MoS2, and Bi2O2Se by a clustering model based on the K-means clustering algorithm, where the white part is the material sample and the black part is the background; Figure 3 As shown in (a) is the result of image segmentation of the material BN, Figure 3 As shown in (b) is the result of image segmentation of the material MoS2, Figure 3 As shown in (c) is the result of image segmentation of the material Bi2O2Se.

[0111] In this embodiment, a fitting experiment was carried out on a batch of samples of experimental materials using the fitted exponential model, with reference to Figure 4 The figure shows the relationship curve between the average gray value of the samples of the experimental material and the predicted average atomic number of the exponential model. The specific points in the figure are the true average atomic numbers of the samples of the experimental material. According to Figure 4 it can be seen that the true average atomic number is close to the predicted average atomic number, and the fitting effect of the exponential model is good.

[0112] Embodiment 2:

[0113] Based on the same inventive concept as Embodiment 1, an average atomic number detection system for a material sample disclosed in an embodiment of the present invention includes:

[0114] An image preprocessing module, configured to: perform image preprocessing on the backscattered electron image of the material sample to be measured obtained, and obtain a processed image;

[0115] An image segmentation module, configured to: use the trained clustering model to perform image segmentation on the processed image, and obtain the area of the material sample to be measured;

[0116] An average gray value calculation module, configured to: calculate the average gray value of the area of the material sample to be measured in the processed image;

[0117] A data fitting module, configured to: input the average gray value of the area of the material sample to be measured in the processed image into the fitted exponential model, and obtain the predicted average atomic number of the material sample to be measured.

[0118] For the specific function implementation of the above modules, refer to the relevant content in the method of Embodiment 1, which will not be elaborated here.

[0119] Embodiment 3:

[0120] This embodiment provides a computer-readable storage medium, on which computer instructions are stored. When the computer instructions are executed by a processor, the steps of the average atomic number detection method for a material sample as described in any one of Embodiment 1 are implemented.

[0121] Embodiment 4:

[0122] This embodiment provides a computer device, including:

[0123] A memory, configured to store computer instructions;

[0124] A processor, configured to execute the computer instructions to implement the steps of the average atomic number detection method for a material sample as described in any one of the first aspect.

[0125] Embodiment 5:

[0126] This embodiment provides a computer program product, including computer instructions, characterized in that when the computer instructions are executed by a processor, the steps of the method for detecting the average atomic number of a material sample described in any one of the first embodiment are implemented.

[0127] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0128] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to the embodiments of the present invention. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0129] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0130] These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one Figure 1 one flow or multiple flows and / or blocks Figure 1 one block or multiple blocks.

[0131] The embodiments of the present invention have been described above in conjunction with the accompanying drawings. However, the present invention is not limited to the above specific embodiments. The above specific embodiments are merely illustrative rather than restrictive. Under the inspiration of the present invention, those of ordinary skill in the art can also make many forms without departing from the spirit of the present invention and the scope protected by the claims. All of these are within the protection scope of the present invention.

Claims

1. A method for detecting the average atomic number of a material sample, characterized in that, Including: Performing image preprocessing on the backscattered electron image of the material sample to be measured obtained, to obtain a processed image; Using the trained clustering model, performing image segmentation on the processed image, to obtain the material sample area to be measured; Calculating the average gray value of the material sample area to be measured in the processed image; Inputting the average gray value of the material sample area to be measured in the processed image into the fitted exponential model, to obtain the predicted average atomic number of the material sample to be measured.

2. The method for detecting the average atomic number of a material sample according to claim 1, wherein The image preprocessing includes gray conversion, normalization processing and Gaussian smoothing processing.

3. The method for detecting the average atomic number of a material sample according to claim 1, wherein The clustering model adopts the K-means clustering algorithm, and the training process of the clustering model includes: Obtaining a training set of backscattered electron images of sample materials; Performing image preprocessing on the training set of backscattered electron images of the sample materials, to obtain a processed sample set; For each sample in the processed sample set, repeating the following training operations until a preset stop condition is met: Optionally selecting two pixel points in the sample as the initial clustering centers of the sample area class and the background area class respectively; Calculating the distances from all pixel points in the sample to the clustering center of the sample area class and the clustering center of the background area class respectively, and assigning each pixel point to the class to which the clustering center with the closest distance belongs; Taking the mean value of all pixel points in the sample area class as the new clustering center of the sample area, and taking the mean value of all pixel points in the background area as the new clustering center of the background area class; After all samples in the processed sample set are trained, a trained clustering model is obtained.

4. The method for detecting the average atomic number of a material sample according to claim 1, characterized in that, Using the trained clustering model, performing image segmentation on the processed image includes: Calculating the distances from each pixel point in the processed image to the clustering center of the sample area class and the clustering center of the background area class respectively, and assigning each pixel point to the class to which the clustering center with the closest distance belongs; Performing image segmentation on the processed image according to all pixel points in the sample area class, to obtain the material sample area.

5. The method for detecting the average atomic number of a material sample according to claim 1, characterized in that, The exponential model uses the nonlinear least squares method for data fitting, and the data fitting process of the exponential model includes: Obtaining a data fitting training set; According to the data fitting training set, establishing an average atomic number set and an average gray value set, and storing the average atomic number of each sample in the data fitting training set and the average gray value of the material sample area of each sample respectively; Initializing the fitting parameters of the exponential model; Repeating the following operations until the sum of squared residuals of the exponential model is less than a threshold, to obtain a fitted exponential model: According to the average atomic number set and the average gray value set, using a nonlinear optimization algorithm to solve the increment of the fitting parameters of the exponential model, and updating the fitting parameters according to the increment of the fitting parameters; According to the average atomic number set and the average gray value set, calculating the sum of squared residuals of the exponential model.

6. The method for detecting the average atomic number of a material sample according to claim 5, characterized in that, The average atomic number of each sample in the data fitting training set is calculated by the following formula: wherein, represents the average atomic number of the sample, represents the number of elements in the sample, represents the mass fraction of the th element in the sample, and represents the atomic number of the The calculation formula of the exponential model is as follows: Among them, represents the predicted average atomic number output by the exponential model, represents the average gray value of the material sample area, represents the fitting parameter of the exponential model; The sum of squared residuals of the exponential model has the following calculation formula: Among them, represents the number of samples in the data fitting training set, represents the average atomic number of the th sample in the data fitting training set, and represents the average gray value of the material sample area of the th sample in the data fitting training set; The update of the fitting parameters is achieved by the following formula: Among them, represents the number of updates of the exponential model, represents the fitting parameter of the exponential model at the th update, represents the fitting parameter of the exponential model at the th update, represents the increment of the fitting parameter of the exponential model.

7. An average atomic number detection system for a material sample, characterized in that, Including: An image preprocessing module, configured to: perform image preprocessing on the backscattered electron image of the material sample to be measured obtained, and obtain a processed image; An image segmentation module, configured to: perform image segmentation on the processed image by using a trained clustering model, and obtain the area of the material sample to be measured; An average gray value calculation module, configured to: calculate the average gray value of the area of the material sample to be measured in the processed image; A data fitting module, configured to: input the average gray value of the area of the material sample to be measured in the processed image into a well-fitted exponential model, and obtain the predicted average atomic number of the material sample to be measured.

8. A computer-readable storage medium having computer instructions stored thereon, characterized in that, When the computer instruction is executed by a processor, the steps of the average atomic number detection method for the material sample according to any one of claims 1-6 are implemented.

9. A computer device, characterized in that, Including: A memory, configured to store computer instructions; A processor, configured to execute the computer instructions to implement the steps of the average atomic number detection method for the material sample according to any one of claims 1-6.

10. A computer program product, comprising computer instructions, characterized in that, When the computer instruction is executed by a processor, the steps of the average atomic number detection method for the material sample according to any one of claims 1-6 are implemented.