SEM image processing method and device, equipment, medium and product
By acquiring the feature information of the SEM image to determine the image type and processing it according to the type, the limitations of processing different types of SEM images in the prior art are solved, and a better image processing effect is achieved.
Patent Information
- Application Number
- CN202411944332.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-26
- Publication Date
- 2025-04-25
AI Technical Summary
The prior art has limitations in processing different types of SEM images, resulting in poor processing effects.
By obtaining the image feature information of the SEM image, including information entropy, energy of the grayscale symbiosis matrix and matrix entropy, and horizontal and vertical distribution profiles, the image type is accurately determined and corresponding image processing strategies are adopted according to the type.
The best quality improvement effect for various types of SEM images is achieved, and the accuracy and effect of image processing are improved.
Smart Images
Figure CN119888323A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of semiconductor integrated circuit technology, and in particular, relates to a SEM image processing method, device, equipment, medium and product. Background Art
[0002] In the semiconductor manufacturing industry, scanning electron microscope (SEM) is an important tool for detecting the microstructure of wafer surface. SEM images provide high-resolution surface morphology information, which is essential for quality control and defect detection in the device manufacturing process. However, due to the inherent characteristics and environmental factors of the SEM imaging process, SEM images usually contain a lot of noise, so SEM images need to be preprocessed to improve image quality.
[0003] In the existing method, SEM images with better quality are obtained by uniformly performing processing steps such as denoising, contrast enhancement and edge sharpening on the SEM images.
[0004] However, due to the diversity of SEM image types, a unified preprocessing method has obvious limitations when processing different types of SEM images, resulting in poor processing effects of SEM images. Summary of the invention
[0005] The embodiments of the present application provide a method, device, equipment, medium and product for processing SEM images, which can improve the processing effect of SEM images.
[0006] In one aspect of an embodiment of the present application, a method for processing a SEM image is provided, comprising:
[0007] Acquire image feature information of the SEM image to be processed, the image feature information including at least one of image information entropy, energy of the image grayscale co-occurrence matrix, matrix entropy of the image grayscale co-occurrence matrix, image lateral distribution profile, and image longitudinal distribution profile;
[0008] Determining the image type of the SEM image to be processed according to the image feature information of the SEM image to be processed;
[0009] The SEM image to be processed is processed by an image processing strategy corresponding to the image type to obtain a processed SEM image.
[0010] In one aspect of an embodiment of the present application, a SEM image processing device is provided, comprising:
[0011] An information acquisition module is used to acquire image feature information of the SEM image to be processed, wherein the image feature information includes at least one of image information entropy, energy of the image grayscale co-occurrence matrix, matrix entropy of the image grayscale co-occurrence matrix, image lateral distribution profile, and image longitudinal distribution profile;
[0012] A type determination module, used to determine the image type of the SEM image to be processed according to the image feature information of the SEM image to be processed;
[0013] The image processing module is used to process the SEM image to be processed by an image processing strategy corresponding to the image type to obtain a processed SEM image.
[0014] In one aspect of an embodiment of the present application, an electronic device is provided, which includes: a memory and a program or instruction stored in the memory and executable on a processor, wherein when the program or instruction is executed by the processor, a method for processing an SEM image as provided in any aspect of the above-mentioned embodiment of the present application is implemented.
[0015] In one aspect of an embodiment of the present application, a readable storage medium is provided, on which a program or instruction is stored. When the program or instruction is executed by a processor, a method for processing an SEM image provided in any aspect of the above-mentioned embodiment of the present application is implemented.
[0016] According to one aspect of an embodiment of the present application, a computer program product is provided. When instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes a method for processing SEM images as provided in any aspect of the above-mentioned embodiment of the present application.
[0017] In the SEM image processing method provided in the embodiment of the present application, the image type of the SEM image to be processed is accurately determined based on the image information entropy of the SEM image to be processed, the energy of the image grayscale co-occurrence matrix, the matrix entropy of the image grayscale co-occurrence matrix, the image lateral distribution profile, and the image longitudinal distribution profile. Then, the SEM image to be processed is processed using a corresponding image processing strategy according to the image type to obtain a processed SEM image. In this way, the present embodiment accurately identifies the image type of the SEM image to be processed and takes corresponding processing measures according to the characteristics of the image type. This enables images of various types to obtain the best quality improvement effect, thereby improving the processing effect of the SEM image. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the technical solution of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.
[0019] Figure 1 It is a schematic flow chart of a method for processing a SEM image provided by an embodiment of the present application;
[0020] Figure 2 It is a schematic diagram of the principle of an image gray level co-occurrence matrix provided by an embodiment of the present application;
[0021] Figure 3 is a schematic diagram of a process for identifying the image type to which a SEM image to be processed belongs, provided by an embodiment of the present application;
[0022] Figure 4 It is a structural schematic diagram of a SEM image processing device provided by an embodiment of the present application;
[0023] Figure 5 It is a schematic diagram of the structure of a SEM image processing device provided in one embodiment of the present application. DETAILED DESCRIPTION
[0024] The features and exemplary embodiments of various aspects of the present application will be described in detail below. In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain the present application, rather than to limit the present application. For those skilled in the art, the present application can be implemented without the need for some of these specific details. The following description of the embodiments is only to provide a better understanding of the present application by illustrating the examples of the present application.
[0025] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the statement "include..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
[0026] It should be noted that the acquisition, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations.
[0027] It should be noted that in the embodiments of the present application, certain software, components, models and other existing solutions in the industry may be mentioned, and they should be regarded as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of the present application, but it does not mean that the applicant has or will necessarily use the solution.
[0028] In the existing methods, SEM images with good quality are obtained by uniformly performing denoising, contrast enhancement, and edge sharpening on the SEM images. However, due to the diversity of SEM images, the unified preprocessing method has obvious limitations when processing different types of SEM images, resulting in poor processing effects of SEM images.
[0029] The purpose of the present application is to provide a method, device, equipment, medium and product for processing SEM images. In the method for processing SEM images provided in the embodiment of the present application, the image type to which the SEM image to be processed belongs is accurately determined based on the image information entropy of the SEM image to be processed, the energy of the image grayscale co-occurrence matrix, the matrix entropy of the image grayscale co-occurrence matrix, the image lateral distribution profile and the image longitudinal distribution profile. Then, the SEM image to be processed is processed according to the image type using a corresponding image processing strategy to obtain a processed SEM image. In this way, the present embodiment accurately identifies the image type of the SEM image to be processed and takes corresponding processing measures according to the characteristics of the image type. This enables images of all types to obtain the best quality improvement effect, thereby improving the processing effect of the SEM image.
[0030] The following describes the specific embodiments of the SEM image processing method, device, equipment, medium and product provided in the embodiments of the present application. First, the SEM image processing method is introduced.
[0031] Figure 1 A schematic flow chart of a method for processing an SEM image is provided. The method for processing an SEM image can be applied to a server. The method for processing an SEM image can include the following steps S101 to S103.
[0032] S101, obtaining image feature information of the SEM image to be processed, the image feature information including at least one of image information entropy, energy of the image grayscale co-occurrence matrix, matrix entropy of the image grayscale co-occurrence matrix, image lateral distribution profile, and image longitudinal distribution profile.
[0033] In this embodiment, image information entropy is an indicator for measuring the information richness of the SEM image, and reflects the complexity and uncertainty of the grayscale distribution in the SEM image.
[0034] The gray-level co-occurrence matrix (GLCM) describes the relationship between the gray values of any two points in the SEM image. The energy of the gray-level co-occurrence matrix reflects the uniformity of the gray-level distribution and the coarseness of the texture of the SEM image. The matrix entropy (GLCM_entropy) of the gray-level co-occurrence matrix reflects the randomness of the distribution of elements in the gray-level co-occurrence matrix and is a measure of the texture complexity of the SEM image.
[0035] The horizontal profile and vertical profile describe the grayscale changes of the SEM image in the horizontal and vertical directions, respectively, and can reflect the edge and texture features of the SEM image. Specifically, the horizontal profile is used to describe the relationship between the grayscale value of each pixel point on a certain horizontal horizontal line in the SEM image and the position; the vertical profile is used to describe the relationship between the grayscale value of each pixel point on a certain vertical vertical line in the SEM image and the position.
[0036] As an example, the server first calculates the probability distribution of each gray level in the SEM image to be processed, and then uses the information entropy formula to determine the image information entropy of the SEM image to be processed; at the same time, the image grayscale co-occurrence matrix of the SEM image to be processed is constructed, so as to determine the energy of the image grayscale co-occurrence matrix and the matrix entropy of the image grayscale co-occurrence matrix according to the distribution of elements in the image grayscale co-occurrence matrix; finally, by calculating the projection of the SEM image to be processed in the horizontal and vertical directions, the distribution curve of the grayscale value is obtained, so as to determine the horizontal distribution profile of the image and the vertical distribution profile of the image.
[0037] S102, determining the image type of the SEM image to be processed according to the image feature information of the SEM image to be processed.
[0038] In this embodiment, the image type to which the SEM image to be processed belongs is determined by comprehensive evaluation of the image information entropy of the SEM image to be processed, the energy of the image grayscale co-occurrence matrix, the matrix entropy of the image grayscale co-occurrence matrix, the image transverse distribution profile, and the image longitudinal distribution profile.
[0039] As an example, the extracted features can be classified using a machine learning or deep learning model (such as a support vector machine, a decision tree, a neural network, etc.).
[0040] Specifically, a training set is first obtained, which contains SEM images of known types and their corresponding image feature information. By training the model, it can learn the mapping relationship between different image types and image feature information. Then, for the SEM image to be processed, its image feature information is extracted and input into the trained model to obtain the image type to which it belongs.
[0041] S103, processing the SEM image to be processed by an image processing strategy corresponding to the image type to obtain a processed SEM image.
[0042] In this embodiment, different image types of SEM images to be processed have corresponding different image processing strategies.
[0043] As an example, the server selects the corresponding image processing strategy according to the image type of the SEM image to be processed. For example, for an image type with more noise, a filtering denoising method can be selected; for an image type with blurred edges, an edge enhancement or sharpening method can be selected; for an image type with complex textures, a texture analysis or segmentation method can be selected.
[0044] In the SEM image processing method provided in this embodiment, the image type of the SEM image to be processed is accurately determined based on the image information entropy of the SEM image to be processed, the energy of the image grayscale co-occurrence matrix, the matrix entropy of the image grayscale co-occurrence matrix, the image lateral distribution profile, and the image longitudinal distribution profile. Then, the corresponding image processing strategy is adopted according to the image type to process the SEM image to be processed to obtain a processed SEM image. In this way, this embodiment accurately identifies the image type of the SEM image to be processed and takes corresponding processing measures according to the characteristics of the image type. This allows various types of images to obtain the best quality improvement effect, thereby improving the processing effect of the SEM image.
[0045] As an optional embodiment, S102 may specifically include:
[0046] Determining the image information entropy of the SEM image to be processed according to the occurrence frequency of each gray value in the SEM image to be processed;
[0047] When the image information entropy is greater than a preset information entropy threshold, determining that the image type of the SEM image to be processed is a type without pattern and pure noise;
[0048] When the image information entropy is less than or equal to the preset information entropy threshold, the image type of the SEM image to be processed is determined according to the energy of the image grayscale co-occurrence matrix, the matrix entropy of the image grayscale co-occurrence matrix, the image transverse distribution profile and the image longitudinal distribution profile.
[0049] In this embodiment, the preset information entropy threshold is used to determine whether the image type of the SEM image to be processed is a type without pattern and pure noise.
[0050] As an example, the server first counts the frequency of occurrence of each gray value in the SEM image to be processed to obtain a gray histogram. According to the gray histogram, the image information entropy of the SEM image to be processed is determined by the following formula 1:
[0051]
[0052] In formula 1, H is used to characterize the image information entropy of the SEM image to be processed, and p i Used to characterize the frequency of occurrence of gray value i, logp i The logarithm used to represent the frequency of occurrence of gray value i, It is used to represent the accumulation of each gray value after calculation.
[0053] Then, the calculated image information entropy of the SEM image to be processed is compared with a preset information entropy threshold. When the image information entropy of the SEM image to be processed is greater than the preset information entropy threshold, it is determined that the SEM image to be processed is a type without pattern and pure noise.
[0054] Then, if the image information entropy is less than or equal to the preset information entropy threshold, it is necessary to further analyze the grayscale co-occurrence matrix, the lateral distribution profile, the longitudinal distribution profile and other features of the SEM image to be processed to determine the image type. Specifically, the image type of the SEM image to be processed can be further determined using a machine learning or deep learning model (such as a support vector machine, a decision tree, a neural network, etc.) based on the grayscale co-occurrence matrix, the lateral distribution profile and the longitudinal distribution profile of the SEM image to be processed.
[0055] Through this embodiment, the image information entropy of the SEM image to be processed is determined according to the frequency of occurrence of each gray value in the SEM image to be processed. Therefore, it is directly determined whether the image type of the SEM image to be processed is a type without pattern and pure noise according to the image information entropy of the SEM image to be processed. In this way, the image type of the SEM image to be processed can be accurately identified without using a large amount of labeled data to train the deep learning model, which can improve the recognition efficiency of the image type.
[0056] As an optional embodiment, determining the image type of the SEM image to be processed according to the energy of the image gray level co-occurrence matrix, the matrix entropy of the image gray level co-occurrence matrix, the image transverse distribution profile, and the image longitudinal distribution profile of the SEM image to be processed may specifically include:
[0057] According to the gray value of each pixel in the SEM image to be processed, an image gray level co-occurrence matrix of the SEM image to be processed is constructed;
[0058] According to the image gray-level co-occurrence matrix, the energy and matrix entropy of the image gray-level co-occurrence matrix are determined;
[0059] When the energy is greater than a preset energy threshold and the matrix entropy is greater than a preset matrix entropy threshold, determining that the image type to which the SEM image to be processed belongs is a type with patterns and large noise;
[0060] When the energy is not greater than a preset energy threshold and / or the matrix entropy is not greater than a preset matrix entropy threshold, the image type of the SEM image to be processed is determined according to the image transverse distribution profile and the image longitudinal distribution profile.
[0061] In this embodiment, the preset energy threshold and the preset matrix entropy threshold are used to determine whether the image type to which the SEM image to be processed belongs is a type with patterns and large noise.
[0062] As an example, the server first pre-sets a pixel spacing and a scanning direction, for example, the preset pixel spacing is 1, and the scanning direction is horizontal scanning or vertical scanning. And according to the pixel spacing and scanning direction, traverse each pixel in the SEM image to be processed to obtain the adjacent pixels of each pixel. Then, according to the grayscale value of the pixel and the grayscale value of its adjacent pixel, a grayscale value pair is formed. The frequency of occurrence of various grayscale value pairs is counted, and the frequency of occurrence of various grayscale value pairs is used as a matrix element to construct an image grayscale co-occurrence matrix of size L×L, where L is the grayscale value level in the SEM image to be processed, that is, the number of different grayscale values contained in one image.
[0063] Then, the energy of the image gray-level co-occurrence matrix is determined by the following formula 2:
[0064] energy=∑ i ∑ j GLCM(i, j) 2 Formula 2
[0065] In formula 2, energy is used to represent the energy of the image gray level co-occurrence matrix, and GLCM(i, j) is used to represent the frequency of occurrence of the gray value pair (i, j) in the image gray level co-occurrence matrix. i ∑ j GLCM(i, j) 2 Used to represent the sum of squares of the frequency of occurrence of each gray value pair in the gray-level co-occurrence matrix of an image.
[0066] Then, the matrix entropy of the image gray-level co-occurrence matrix is determined by the following formula 3:
[0067] GLCM-entropy=-∑ i ∑ jGLCM(i,j)* logGLCM(i,j) Formula 3
[0068] In formula 3, GLCM(i, j) is used to represent the frequency of occurrence of gray value pairs (i, j) in the gray level co-occurrence matrix of the image, -∑ i ∑ j GLCM(i, j)*logGLCM(i, j) is used to represent the negative sum of the product of the occurrence frequency of each gray value and the logarithm of the occurrence frequency in the image gray level co-occurrence matrix, and GLCM-entropy is used to represent the matrix entropy of the image gray level co-occurrence matrix.
[0069] Then, the calculated energy and matrix entropy are compared with the preset energy threshold and the preset matrix entropy threshold, respectively. When the energy is greater than the preset energy threshold and the matrix entropy is greater than the preset matrix entropy threshold, it is determined that the image type of the SEM image to be processed is a type with patterns and large noise;
[0070] Then, if the energy is not greater than the preset energy threshold, and / or the matrix entropy is not greater than the preset matrix entropy threshold, the image lateral distribution profile and the image longitudinal distribution profile of the SEM image to be processed are further calculated, and the image type to which the SEM image to be processed belongs is determined based on the size relationship between the image lateral distribution profile and the image longitudinal distribution profile.
[0071] Through this embodiment, the energy and matrix entropy of the image gray-level co-occurrence matrix are determined according to the image gray-level co-occurrence matrix. Therefore, it is directly determined whether the image type to which the SEM image to be processed belongs is a type with patterns and large noise according to the energy and matrix entropy of the image gray-level co-occurrence matrix. In this way, the image type to which the SEM image to be processed belongs can be accurately identified without using a large amount of labeled data to train the deep learning model, which can improve the recognition efficiency of the image type.
[0072] As an optional embodiment, constructing an image grayscale co-occurrence matrix of the SEM image to be processed according to the grayscale value of each pixel in the SEM image to be processed may specifically include:
[0073] Get pixel spacing and scanning direction;
[0074] Based on the pixel spacing and the scanning direction, various types of gray value pairs included in the SEM image to be processed are determined, wherein the gray value pairs include gray values of two adjacent pixel points;
[0075] Taking each gray value pair as a search condition, traversing the SEM image to be processed, and determining the occurrence frequency of each gray value pair in the SEM image to be processed;
[0076] The occurrence frequency of each gray value pair is used as a matrix element to construct the image gray level co-occurrence matrix of the SEM image to be processed.
[0077] As an example, Figure 2 As shown in the figure, a schematic diagram of the principle of an image grayscale co-occurrence matrix is provided. The left side shows the grayscale value of each pixel included in the SEM image to be processed, and the right side shows the image grayscale co-occurrence matrix corresponding to the SEM image to be processed. According to the grayscale value of each pixel included in the SEM image to be processed, it can be seen that the SEM image to be processed includes 9 different grayscale values, so the size of the image grayscale co-occurrence matrix is 9×9.
[0078] Among them, the preset pixel spacing is 1, and the scanning direction is horizontal scanning, then (2,3) constitutes a gray value pair, and (3,5) constitutes a gray value pair. In addition, the gray value pair has directionality, that is, (2,3) and (3,2) belong to different gray value pairs. In this way, the server determines the various types of gray value pairs included in the SEM image to be processed based on the pixel spacing of 1 and the scanning direction of horizontal scanning. Then, using each gray value pair as the search condition, traverse the SEM image to be processed to determine the frequency of occurrence of each gray value pair in the SEM image to be processed. As can be seen from the figure, the frequency of occurrence of the gray value pair (2,3) is 1, and the frequency of occurrence of the gray value pair (3,5) is 3.
[0079] Then, the image grayscale co-occurrence matrix of the SEM image to be processed is constructed with the occurrence frequency of various grayscale value pairs as matrix elements. Specifically, the first grayscale value in the grayscale value pair corresponds to the row of the image grayscale co-occurrence matrix, and the second grayscale value in the grayscale value pair corresponds to the column of the image grayscale co-occurrence matrix. Then the matrix element value of the 2nd row and 3rd column in the image grayscale co-occurrence matrix is 1, and the matrix element value of the 3rd row and 5th column in the image grayscale co-occurrence matrix is 3.
[0080] Through this embodiment, various types of gray value pairs included in the SEM image to be processed are determined based on the pixel pitch and the scanning direction. Thus, the image gray level co-occurrence matrix of the SEM image to be processed is constructed according to the frequency of occurrence of each gray level pair. In this way, the image gray level co-occurrence matrix of the SEM image to be processed can be accurately constructed, which helps to identify the image type of the SEM image to be processed according to the image gray level co-occurrence matrix.
[0081] As an optional embodiment, determining the image type of the SEM image to be processed according to the transverse distribution profile of the image and the longitudinal distribution profile of the image may specifically include:
[0082] Obtaining adjacent grayscale values of each grayscale value, wherein the adjacent grayscale values include a first adjacent grayscale value that is N greater than the grayscale value and a second adjacent grayscale value that is N less than the grayscale value, where N is a preset positive integer;
[0083] Determine, according to the occurrence frequency of each gray value pair in the SEM image to be processed, a first image lateral distribution profile and a first image longitudinal distribution profile of a first adjacent gray value, and a second image lateral distribution profile and a second image longitudinal distribution profile of a second adjacent gray value;
[0084] Determine a gradient value sequence of a lateral distribution profile of the SEM image to be processed according to the lateral distribution profile of the first image and the lateral distribution profile of the second image corresponding to each gray value, and determine a gradient value sequence of a longitudinal distribution profile of the SEM image to be processed according to the longitudinal distribution profile of the first image and the longitudinal distribution profile of the second image corresponding to each gray value;
[0085] The image type of the SEM image to be processed is determined according to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence.
[0086] In this embodiment, the lateral distribution profile of the image can be specifically determined by the following formula 4:
[0087]
[0088] In formula 4, horizontal profile(i) It is used to characterize the lateral distribution profile of the image with gray value i, N is used to characterize the number of gray value types included in the SEM image to be processed, and f(i, j) is used to characterize the frequency of occurrence of the gray value pair (i, j), that is, the frequency of occurrence of the gray value pair (i, j) divided by the sum of the frequency of occurrence of each gray value in the image grayscale co-occurrence matrix.
[0089] The vertical distribution profile of the image can be specifically determined by the following formula 5:
[0090]
[0091] In formula 5, vertical profile(j) It is used to characterize the longitudinal distribution profile of the image with gray value j, N is used to characterize the number of gray value types included in the SEM image to be processed, and f(i, j) is used to characterize the frequency of occurrence of the gray value pair (i, j), that is, the frequency of occurrence of the gray value pair (i, j) divided by the sum of the frequency of occurrence of each gray value in the image gray co-occurrence matrix.
[0092] As an example, the adjacent grayscale values corresponding to the preset grayscale value include a first adjacent grayscale value that is 1 greater than the grayscale value and a second adjacent grayscale value that is 1 less than the grayscale value.
[0093] The distribution profile gradient value of each gray value can be specifically determined by the following formula 6:
[0094]
[0095] In formula 6, V i Used to characterize the distribution profile gradient value of gray value i, profile(i+1) is used to characterize the image distribution profile of the first adjacent gray value of gray value i, and profile(i-1) is used to characterize the image distribution profile of the second adjacent gray value of gray value i.
[0096] Specifically, when it is necessary to calculate the lateral distribution contour gradient value of grayscale value i, the lateral distribution contour of the image of the first adjacent grayscale value of grayscale value i is subtracted from the lateral distribution contour of the image of the second adjacent grayscale value of grayscale value i, and then divided by 2 to obtain the lateral distribution contour gradient value of grayscale value i.
[0097] When it is necessary to calculate the gradient value of the longitudinal distribution profile of gray value i, the longitudinal distribution profile of the image of the first adjacent gray value of gray value i is subtracted from the longitudinal distribution profile of the image of the second adjacent gray value of gray value i, and then divided by 2 to obtain the gradient value of the longitudinal distribution profile of gray value i.
[0098] In this way, the lateral distribution profile gradient value and the longitudinal distribution profile gradient value of each gray value in the SEM image to be processed can be calculated by the above formula 6.
[0099] Then, the lateral distribution profile gradient values of each gray value in the SEM image to be processed are arranged in order from large to small to obtain a lateral distribution profile gradient value sequence of the SEM image to be processed; at the same time, the longitudinal distribution profile gradient values of each gray value in the SEM image to be processed are arranged in order from large to small to obtain a longitudinal distribution profile gradient value sequence of the SEM image to be processed.
[0100] Finally, the image type to which the SEM image to be processed belongs is determined according to each lateral distribution profile gradient value in the lateral distribution profile gradient value sequence and each longitudinal distribution profile gradient value in the longitudinal distribution profile gradient value sequence.
[0101] Through this embodiment, according to the occurrence frequency of each gray value pair in the SEM image to be processed, the image lateral distribution profile and the image longitudinal distribution profile of each gray value are determined, so that the lateral distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence of the SEM image to be processed can be accurately determined. In this way, the image type to which the SEM image to be processed belongs can be directly identified based on the lateral distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence, which can improve the recognition efficiency of the image type.
[0102] As an optional embodiment, determining the image type of the SEM image to be processed according to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence may specifically include:
[0103] According to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence, determining the largest first Z transverse distribution profile gradient values in the transverse distribution profile gradient value sequence and the largest first Z longitudinal distribution profile gradient values in the longitudinal distribution profile gradient value sequence, wherein Z is a preset positive integer;
[0104] When the sum of the maximum first Z transverse distribution profile gradient values is greater than the preset gradient value threshold, or the sum of the maximum first Z longitudinal distribution profile gradient values is greater than the preset gradient value threshold, determining that the image type of the SEM image to be processed is a type without pattern and pure noise;
[0105] When the sum of the largest first Z transverse distribution profile gradient values is not greater than the preset gradient value threshold, and the sum of the largest first Z longitudinal distribution profile gradient values is not greater than the preset gradient value threshold, it is determined that the image type to which the SEM image to be processed belongs is a type with pattern and low noise.
[0106] In this embodiment, the server obtains the largest first Z transverse distribution profile gradient values from the transverse distribution profile gradient value sequence, and simultaneously obtains the largest first Z longitudinal distribution profile gradient values from the longitudinal distribution profile gradient value sequence.
[0107] Then, the sum of the first Z largest transverse distribution profile gradient values in the transverse distribution profile gradient value sequence and the sum of the first Z largest longitudinal distribution profile gradient values in the longitudinal distribution profile gradient value sequence are calculated respectively.
[0108] Finally, the sum of the largest first Z transverse distribution profile gradient values and the sum of the largest first Z longitudinal distribution profile gradient values are compared with the preset gradient value thresholds. If the sum of the largest first Z transverse distribution profile gradient values is greater than the preset gradient value threshold, or the sum of the largest first Z longitudinal distribution profile gradient values is greater than the preset gradient value threshold, the SEM image to be processed is determined to be of the type of "no pattern and pure noise"; if the sum of the largest first Z transverse distribution profile gradient values is not greater than the preset gradient value threshold, and the sum of the largest first Z longitudinal distribution profile gradient values is not greater than the preset gradient value threshold, the SEM image to be processed is determined to be of the type of "with pattern and low noise".
[0109] Through this embodiment, the image type of the SEM image to be processed can be identified based on the largest first Z transverse distribution profile gradient values in the transverse distribution profile gradient value sequence and the largest first Z longitudinal distribution profile gradient values in the longitudinal distribution profile gradient value sequence. It is not necessary to make a judgment based on all the distribution profile gradient values in the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence, thereby reducing the amount of calculation and improving the efficiency of image type recognition.
[0110] As a scenario example, Figure 3 As shown, a flow chart of identifying the image type to which the SEM image to be processed belongs is provided. First, the server calculates the image information entropy of the SEM image to be processed through S301. Then, it is determined through S302 whether the image information entropy of the SEM image to be processed is greater than a preset information entropy threshold. In the case where the image information entropy is greater than the preset information entropy threshold, the image type to which the SEM image to be processed belongs is determined as a type without pattern and pure noise through S303.
[0111] If the image information entropy is not greater than the preset information entropy threshold, the energy and matrix entropy of the image grayscale co-occurrence matrix are calculated by S304. Then, it is determined by S305 whether the energy is greater than the preset energy threshold and whether the matrix entropy is greater than the preset matrix threshold. If the energy is greater than the preset energy threshold and the matrix entropy is greater than the preset matrix threshold, it is determined by S306 that the image type of the SEM image to be processed is a type with patterns and large noise.
[0112] When the energy is not greater than the preset energy threshold or the matrix entropy is not greater than the preset matrix threshold, the sum of the largest first Z transverse distribution profile gradient values in the transverse distribution profile gradient value sequence and the sum of the largest first Z longitudinal distribution profile gradient values in the longitudinal distribution profile gradient value sequence are calculated by S307. Then, it is determined by S308 whether the sum of the largest first Z transverse distribution profile gradient values or the sum of the largest first Z longitudinal distribution profile gradient values is greater than the preset gradient value threshold.
[0113] When the sum of the largest first Z lateral distribution profile gradient values or the sum of the largest first Z longitudinal distribution profile gradient values is greater than the preset gradient value threshold, S309 determines that the image type of the SEM image to be processed is a type without pattern and pure noise; when the sum of the largest first Z lateral distribution profile gradient values and the sum of the largest first Z longitudinal distribution profile gradient values are not greater than the preset gradient value threshold, S310 determines that the image type of the SEM image to be processed is a type with pattern and low noise.
[0114] As an optional embodiment, the image type is a type without pattern and pure noise;
[0115] S103 may specifically include:
[0116] The image format conversion operation is performed on the SEM image to be processed to obtain a processed SEM image.
[0117] In this embodiment, the type of no pattern and pure noise is used to characterize the SEM image which has no obvious pattern or structure and is mainly composed of noise.
[0118] As an example, when the SEM image to be processed is of the type without pattern and pure noise, the server first reads the SEM image to be processed, assuming that the SEM image to be processed is stored in the form of a 16-bit grayscale image, that is, its grayscale value range is between 0 and 65535.
[0119] Then, the SEM image to be processed is converted into an 8-bit grayscale image for storage, that is, the grayscale value range of the SEM image to be processed is mapped from 0-65535 to 0-255. Specifically, this can be achieved by linear scaling, that is, dividing the grayscale value of each pixel by 256.
[0120] Through this embodiment, for the unprocessed SEM image of the type without pattern and pure noise, the unprocessed SEM image is only converted into an 8-bit grayscale image for storage, and no other operations are required. In this way, it is possible to avoid performing too many meaningless processing operations on the unprocessed SEM image of the type without pattern and pure noise, thereby improving the processing efficiency of the unprocessed SEM image of the type without pattern and pure noise.
[0121] As an optional embodiment, the image type is a type with patterns and little noise;
[0122] S103 may specifically include:
[0123] Perform filtering and denoising on the SEM image to be processed to obtain a denoised SEM image;
[0124] Performing a sharpening operation on the denoised SEM image to obtain a sharpened SEM image;
[0125] The sharpened SEM image is linearly stretched and the image format is converted to obtain a processed SEM image.
[0126] In this embodiment, the patterned and low-noise type is used to characterize SEM images with obvious patterns or structures.
[0127] As an example, when the SEM image to be processed is of a type with patterns and little noise, the server first selects a filter, such as a median filter, and then applies the selected median filter to the SEM image to be processed, and denoises the SEM image to be processed by median filtering to obtain a denoised SEM image.
[0128] Then, a sharpening operation is performed on the denoised SEM image using a sharpening filter to enhance the edge and details of the denoised SEM image to obtain a sharpened SEM image. For example, an Unsharp Mask Sharpening (USM) operation can be performed on the denoised SEM image.
[0129] Finally, for the sharpened SEM image, the linear stretching technique was used to adjust the grayscale range of the image to enhance the contrast, and the image was converted into an 8-bit grayscale image for storage.
[0130] Through this embodiment, filtering and denoising, sharpening, linear stretching and image format conversion operations are performed on the SEM images to be processed that have patterns and little noise, so that the processing quality of the SEM images to be processed that have patterns and little noise can be improved.
[0131] As an optional embodiment, the image type is a type with patterns and large noise;
[0132] S103 may specifically include:
[0133] Obtaining a grayscale histogram of the SEM image to be processed;
[0134] According to the grayscale histogram, each grayscale value whose distance from the grayscale mean value of the SEM image to be processed is greater than a preset grayscale threshold is screened out to obtain a screened SEM image;
[0135] Filter and denoise the screened SEM image to obtain a target intermediate image;
[0136] The target intermediate image is linearly stretched and the image format is converted to obtain a processed SEM image.
[0137] In this example, the patterned and noisy type is used to characterize SEM images that contain useful patterns or structures but have a high noise level.
[0138] As an example, when the SEM image to be processed is of a type with patterns and large noise. The server first obtains the grayscale histogram of the SEM image to be processed, and determines the grayscale mean of the SEM image to be processed according to the grayscale histogram. Then, according to the normal distribution of the grayscale histogram, a grayscale value interval is constructed with three standard deviations (sigma) above and below the grayscale mean as the range. The grayscale values in the SEM image to be processed that are within the grayscale value interval are retained, and the grayscale values in the SEM image to be processed that are not within the grayscale value interval are screened out, thereby obtaining a screened SEM image.
[0139] Then, a filter, such as a Gaussian filter, is selected, and the selected median filter is applied to the SEM image after screening, and the SEM image after screening is denoised by Gaussian filtering to obtain a target intermediate image.
[0140] Finally, for the target intermediate image, the linear stretching technique is used to adjust the grayscale range of the image to enhance the contrast, and the image is converted into an 8-bit grayscale image for storage.
[0141] Through this embodiment, for the SEM image with patterns and large noise to be processed, gray value screening, filtering and denoising, linear stretching and image format conversion operations are performed respectively. In this way, the processing quality of the SEM image with patterns and large noise to be processed can be improved.
[0142] Processing method based on SEM images. Accordingly, the present application also provides a specific embodiment of a processing device for SEM images.
[0143] like Figure 4 As shown, the SEM image processing device 400 provided in the embodiment of the present application includes an information acquisition module 410 , a type determination module 420 and an image processing module 430 .
[0144] The information acquisition module 410 is used to obtain image feature information of the SEM image to be processed, and the image feature information includes at least one of image information entropy, energy of the image grayscale co-occurrence matrix, matrix entropy of the image grayscale co-occurrence matrix, image lateral distribution profile, and image longitudinal distribution profile.
[0145] The type determination module 420 is used to determine the image type of the SEM image to be processed according to the image feature information of the SEM image to be processed.
[0146] The image processing module 430 is used to process the SEM image to be processed by using an image processing strategy corresponding to the image type to obtain a processed SEM image.
[0147] As an optional embodiment, the type determination module 420 specifically includes the following submodules:
[0148] An information entropy determination submodule, used to determine the image information entropy of the SEM image to be processed according to the occurrence frequency of each gray value in the SEM image to be processed;
[0149] A type determination submodule, used for determining that the image type of the SEM image to be processed is a type without pattern and pure noise when the image information entropy is greater than a preset information entropy threshold;
[0150] The type determination submodule is also used to determine the image type of the SEM image to be processed based on the energy of the image grayscale co-occurrence matrix, the matrix entropy of the image grayscale co-occurrence matrix, the image transverse distribution profile and the image longitudinal distribution profile of the SEM image to be processed when the image information entropy is less than or equal to a preset information entropy threshold.
[0151] As an optional embodiment, the type determination submodule specifically includes the following units:
[0152] A matrix construction unit, used for constructing an image grayscale co-occurrence matrix of the SEM image to be processed according to the grayscale value of each pixel point in the SEM image to be processed;
[0153] An information acquisition unit, used for determining the energy and matrix entropy of the image gray-level co-occurrence matrix according to the image gray-level co-occurrence matrix;
[0154] A type determination unit, used for determining that the image type of the SEM image to be processed is a type with patterns and large noise when the energy is greater than a preset energy threshold and the matrix entropy is greater than a preset matrix entropy threshold;
[0155] The type determination unit is also used to determine the image type of the SEM image to be processed according to the horizontal distribution profile and the vertical distribution profile of the image when the energy is not greater than a preset energy threshold and / or the matrix entropy is not greater than a preset matrix entropy threshold.
[0156] As an optional embodiment, the matrix construction unit is specifically used for:
[0157] Get pixel spacing and scanning direction;
[0158] Based on the pixel spacing and the scanning direction, various types of gray value pairs included in the SEM image to be processed are determined, wherein the gray value pairs include gray values of two adjacent pixel points;
[0159] Taking each gray value pair as a search condition, traversing the SEM image to be processed, and determining the occurrence frequency of each gray value pair in the SEM image to be processed;
[0160] The occurrence frequency of each gray value pair is used as a matrix element to construct the image gray level co-occurrence matrix of the SEM image to be processed.
[0161] As an optional embodiment, the type determination unit specifically includes the following subunits:
[0162] A grayscale value acquisition subunit, used to acquire adjacent grayscale values of each grayscale value, wherein the adjacent grayscale values include a first adjacent grayscale value greater than the grayscale value by N and a second adjacent grayscale value less than the grayscale value by N, where N is a preset positive integer;
[0163] A profile determination subunit, for determining a first image lateral distribution profile and a first image longitudinal distribution profile of a first adjacent gray value, and a second image lateral distribution profile and a second image longitudinal distribution profile of a second adjacent gray value according to the occurrence frequency of each gray value pair in the SEM image to be processed;
[0164] a gradient value determination subunit, for determining a gradient value sequence of a lateral distribution profile of the SEM image to be processed according to the lateral distribution profile of the first image and the lateral distribution profile of the second image corresponding to each gray value, and for determining a gradient value sequence of a longitudinal distribution profile of the SEM image to be processed according to the longitudinal distribution profile of the first image and the longitudinal distribution profile of the second image corresponding to each gray value;
[0165] The type determination subunit is used to determine the image type to which the SEM image to be processed belongs according to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence.
[0166] As an optional embodiment, the type determination subunit is specifically used for:
[0167] According to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence, determining the largest first Z transverse distribution profile gradient values in the transverse distribution profile gradient value sequence and the largest first Z longitudinal distribution profile gradient values in the longitudinal distribution profile gradient value sequence, where Z is a preset positive integer;
[0168] When the sum of the maximum first Z transverse distribution profile gradient values is greater than the preset gradient value threshold, or the sum of the maximum first Z longitudinal distribution profile gradient values is greater than the preset gradient value threshold, determining that the image type of the SEM image to be processed is a type without pattern and pure noise;
[0169] When the sum of the largest first Z transverse distribution profile gradient values is not greater than the preset gradient value threshold, and the sum of the largest first Z longitudinal distribution profile gradient values is not greater than the preset gradient value threshold, it is determined that the image type to which the SEM image to be processed belongs is a type with pattern and low noise.
[0170] As an optional embodiment, the image type is a type without pattern and pure noise;
[0171] The image processing module 430 is specifically used for:
[0172] The image format conversion operation is performed on the SEM image to be processed to obtain a processed SEM image.
[0173] As an optional embodiment, the image type is a type with patterns and little noise;
[0174] The image processing module 430 is specifically used for:
[0175] Perform filtering and denoising on the SEM image to be processed to obtain a denoised SEM image;
[0176] Performing a sharpening operation on the denoised SEM image to obtain a sharpened SEM image;
[0177] The sharpened SEM image is linearly stretched and the image format is converted to obtain a processed SEM image.
[0178] As an optional embodiment, the image type is a type with patterns and large noise;
[0179] The image processing module 430 is specifically used for:
[0180] Obtaining a grayscale histogram of the SEM image to be processed;
[0181] According to the grayscale histogram, each grayscale value whose distance from the grayscale mean value of the SEM image to be processed is greater than a preset grayscale threshold is screened out to obtain a screened SEM image;
[0182] Filter and denoise the screened SEM image to obtain a target intermediate image;
[0183] The target intermediate image is linearly stretched and the image format is converted to obtain a processed SEM image.
[0184] Processing method based on SEM image. Accordingly, the present application also provides a specific embodiment of a processing device for SEM image.
[0185] Figure 5 A schematic diagram of the hardware structure of a SEM image processing device provided in an embodiment of the present application is shown.
[0186] The SEM image processing device may include a processor 501 and a memory 502 storing computer program instructions.
[0187] Specifically, the processor 501 may include a central processing unit (CPU), or an application specific integrated circuit (ASIC), or may be configured to implement one or more integrated circuits of the embodiments of the present application.
[0188] The memory 502 may include a large capacity memory for data or instructions. By way of example and not limitation, the memory 502 may include a hard disk drive (HDD), a floppy disk drive, a flash memory, an optical disk, a magneto-optical disk, a magnetic tape, or a universal serial bus (USB) drive or a combination of two or more of these. In appropriate cases, the memory 502 may include a removable or non-removable (or fixed) medium. In appropriate cases, the memory 502 may be inside or outside the integrated gateway disaster recovery device. In a specific embodiment, the memory 502 is a non-volatile solid-state memory.
[0189] The processor 501 reads and executes the computer program instructions stored in the memory 502 to implement any one of the SEM image processing methods in the above embodiments.
[0190] In one example, the SEM image processing device may further include a communication interface 503 and a bus 510. Figure 5 As shown, the processor 501, the memory 502, and the communication interface 503 are connected via a bus 510 and communicate with each other.
[0191] The communication interface 503 is mainly used to implement communication between various modules, devices, units and / or equipment in the embodiments of the present application.
[0192] The bus 510 includes hardware, software, or both, coupling the components of the SEM image processing device to each other. By way of example and not limitation, the bus may include an accelerated graphics port (AGP) or other graphics bus, an enhanced industry standard architecture (EISA) bus, a front-side bus (FSB), a hypertransport (HT) interconnect, an industry standard architecture (ISA) bus, an infinite bandwidth interconnect, a low pin count (LPC) bus, a memory bus, a microchannel architecture (MCA) bus, a peripheral component interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a serial advanced technology attachment (SATA) bus, a video electronics standard association local (VLB) bus, or other suitable buses or a combination of two or more of these. Where appropriate, the bus 510 may include one or more buses. Although the embodiments of the present application describe and illustrate specific buses, the present application contemplates any suitable bus or interconnect.
[0193] In addition, in combination with the SEM image processing method in the above embodiment, the embodiment of the present application can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when the computer program instructions are executed by the processor, any one of the SEM image processing methods in the above embodiment is implemented.
[0194] In addition, in combination with the SEM image processing method in the above-mentioned embodiment, the embodiment of the present application may provide a computer program product for implementation. When the instructions in the computer program product are executed by the processor of an electronic device, the electronic device executes the SEM image processing method provided in any aspect of the above-mentioned embodiment of the present application.
[0195] It should be clear that the present application is not limited to the specific configuration and processing described above and shown in the figures. For the sake of simplicity, a detailed description of the known method is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of the present application is not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between the steps after understanding the spirit of the present application.
[0196] The functional blocks shown in the above-described block diagram can be implemented as hardware, software, firmware or a combination thereof. When implemented in hardware, it can be, for example, an electronic circuit, an application specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc. When implemented in software, the elements of the present application are programs or code segments that are used to perform the required tasks. The program or code segment can be stored in a machine-readable medium, or transmitted on a transmission medium or a communication link by a data signal carried in a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, optical fiber media, radio frequency (RF) links, etc. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.
[0197] It should also be noted that the exemplary embodiments mentioned in this application describe some methods or systems based on a series of steps or devices. However, this application is not limited to the order of the above steps, that is, the steps can be performed in the order mentioned in the embodiment, or in a different order from the embodiment, or several steps can be performed simultaneously.
[0198] Aspects of the present disclosure are described above with reference to the flowchart and / or block diagram of the method, device (system) and computer program product according to the embodiment of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device to produce a machine so that these instructions executed by the processor of the computer or other programmable data processing device enable the implementation of the function / action specified in one or more boxes of the flowchart and / or block diagram. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field programmable logic circuit. It can also be understood that each box in the block diagram and / or flowchart and the combination of boxes in the block diagram and / or flowchart can also be implemented by dedicated hardware that performs a specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions.
[0199] The above is only a specific implementation of the present application. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, modules and units described above can refer to the corresponding processes in the aforementioned method embodiments, and will not be repeated here. It should be understood that the protection scope of the present application is not limited to this. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in this application, and these modifications or replacements should be included in the protection scope of this application.
Claims
1. A method for processing a SEM image, characterized in that: include: Acquire image feature information of the SEM image to be processed, wherein the image feature information includes at least one of image information entropy, energy of the image grayscale co-occurrence matrix, matrix entropy of the image grayscale co-occurrence matrix, image lateral distribution profile, and image longitudinal distribution profile; Determining the image type of the SEM image to be processed according to the image feature information of the SEM image to be processed; The SEM image to be processed is processed by an image processing strategy corresponding to the image type to obtain a processed SEM image.
2. The method according to claim 1, characterized in that The step of determining the image type of the SEM image to be processed according to the image feature information of the SEM image to be processed includes: Determining the image information entropy of the SEM image to be processed according to the occurrence frequency of each gray value in the SEM image to be processed; When the image information entropy is greater than a preset information entropy threshold, determining that the image type to which the SEM image to be processed belongs is a type without pattern and pure noise; When the image information entropy is less than or equal to the preset information entropy threshold, the image type of the SEM image to be processed is determined based on the energy of the image grayscale co-occurrence matrix of the SEM image to be processed, the matrix entropy of the image grayscale co-occurrence matrix, the image transverse distribution profile, and the image longitudinal distribution profile.
3. The method according to claim 2, characterized in that The step of determining the image type of the SEM image to be processed according to the energy of the image grayscale co-occurrence matrix of the SEM image to be processed, the matrix entropy of the image grayscale co-occurrence matrix, the image transverse distribution profile, and the image longitudinal distribution profile comprises: Constructing the image grayscale co-occurrence matrix of the SEM image to be processed according to the grayscale value of each pixel in the SEM image to be processed; Determining the energy and matrix entropy of the image gray level co-occurrence matrix according to the image gray level co-occurrence matrix; When the energy is greater than a preset energy threshold and the matrix entropy is greater than a preset matrix entropy threshold, determining that the image type to which the SEM image to be processed belongs is a type with patterns and large noise; When the energy is not greater than the preset energy threshold and / or the matrix entropy is not greater than the preset matrix entropy threshold, the image type of the SEM image to be processed is determined according to the transverse distribution profile of the image and the longitudinal distribution profile of the image.
4. The method according to claim 3, characterized in that The step of constructing the image grayscale co-occurrence matrix of the SEM image to be processed according to the grayscale value of each pixel in the SEM image to be processed comprises: Get pixel spacing and scanning direction; Based on the pixel pitch and the scanning direction, determining various types of gray value pairs included in the SEM image to be processed, wherein the gray value pairs include gray values of two adjacent pixel points; Taking each of the gray value pairs as a search condition, traversing the SEM image to be processed, and determining the occurrence frequency of each of the gray value pairs in the SEM image to be processed; The image grayscale co-occurrence matrix of the SEM image to be processed is constructed by taking the occurrence frequency of each grayscale value pair as a matrix element.
5. The method according to claim 3, characterized in that: Determining the image type of the SEM image to be processed according to the transverse distribution profile of the image and the longitudinal distribution profile of the image includes: Obtaining adjacent grayscale values of each grayscale value, wherein the adjacent grayscale values include a first adjacent grayscale value that is N greater than the grayscale value and a second adjacent grayscale value that is N less than the grayscale value, where N is a preset positive integer; Determine, according to the occurrence frequency of each gray value pair in the SEM image to be processed, a first image lateral distribution profile and a first image longitudinal distribution profile of the first adjacent gray values, and a second image lateral distribution profile and a second image longitudinal distribution profile of the second adjacent gray values; Determine a gradient value sequence of the lateral distribution profile of the SEM image to be processed according to the lateral distribution profile of the first image and the lateral distribution profile of the second image corresponding to each of the grayscale values, and determine a gradient value sequence of the longitudinal distribution profile of the SEM image to be processed according to the longitudinal distribution profile of the first image and the longitudinal distribution profile of the second image corresponding to each of the grayscale values; The image type to which the SEM image to be processed belongs is determined according to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence.
6. The method according to claim 5, characterized in that The step of determining the image type of the SEM image to be processed according to the lateral distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence includes: According to the transverse distribution profile gradient value sequence and the longitudinal distribution profile gradient value sequence, determining the largest first Z transverse distribution profile gradient values in the transverse distribution profile gradient value sequence and the largest first Z longitudinal distribution profile gradient values in the longitudinal distribution profile gradient value sequence, where Z is a preset positive integer; When the sum of the maximum first Z transverse distribution profile gradient values is greater than the preset gradient value threshold, or the sum of the maximum first Z longitudinal distribution profile gradient values is greater than the preset gradient value threshold, determining that the image type to which the SEM image to be processed belongs is a type without pattern and pure noise; When the sum of the largest first Z transverse distribution profile gradient values is not greater than the preset gradient value threshold, and the sum of the largest first Z longitudinal distribution profile gradient values is not greater than the preset gradient value threshold, it is determined that the image type to which the SEM image to be processed belongs is a type with pattern and low noise.
7. The method according to any one of claims 1 to 6, characterized in that: The image type is a type without pattern and pure noise; The step of processing the SEM image to be processed by using an image processing strategy corresponding to the image type to obtain a processed SEM image comprises: An image format conversion operation is performed on the SEM image to be processed to obtain a processed SEM image.
8. The method according to any one of claims 1 to 6, characterized in that: The image type is a type with patterns and low noise; The step of processing the SEM image to be processed by using an image processing strategy corresponding to the image type to obtain a processed SEM image comprises: Performing filtering and denoising on the SEM image to be processed to obtain a denoised SEM image; Performing a sharpening operation on the denoised SEM image to obtain a sharpened SEM image; The sharpened SEM image is linearly stretched and image format converted to obtain a processed SEM image.
9. The method according to any one of claims 1 to 6, characterized in that: The image type is a type with patterns and large noise; The step of processing the SEM image to be processed by using an image processing strategy corresponding to the image type to obtain a processed SEM image comprises: Obtaining a grayscale histogram of the SEM image to be processed; According to the grayscale histogram, each grayscale value whose distance from the grayscale mean value of the SEM image to be processed is greater than a preset grayscale threshold is screened out to obtain a screened SEM image; Filtering and denoising the screened SEM image to obtain a target intermediate image; The target intermediate image is linearly stretched and image format converted to obtain a processed SEM image.
10. An electronic device, characterized in that: The device comprises: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, the SEM image processing method according to any one of claims 1 to 9 is implemented.
11. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer program instructions, and when the computer program instructions are executed by a processor, the method for processing the SEM image according to any one of claims 1 to 9 is implemented.
12. A computer program product, characterized in that When the instructions in the computer program product are executed by a processor of an electronic device, the electronic device executes the SEM image processing method according to any one of claims 1 to 9.