Scanning electron microscope resolving power evaluation method and device, computer equipment and storage medium

Abstract

The embodiment of the invention provides a scanning electron microscope resolving power evaluation method, computer equipment and a storage medium. The method comprises the following steps: acquiring a scanning image of a to-be-evaluated scanning electron microscope on a sample; performing image enhancement processing on the scanning image by using an image enhancement model to obtain an enhanced scanning image; and evaluating the resolving power of the scanning electron microscope to be evaluated through the enhanced scanning image. According to the method, on one hand, a sample is adopted as a scanning object of the scanning electron microscope to be evaluated, and the accuracy of an evaluation result is improved by utilizing the material property of a mesoporous molecular sieve, and on the other hand, the resolution of the scanning electron microscope to be evaluated is evaluated through an enhanced scanning image. Compared with the scanning image before enhancement, the enhanced scanning image has higher image quality, so that the accuracy of the evaluation result can be obviously improved by evaluating the resolving power of the scanning electron microscope to be evaluated through the enhanced scanning image.

Description

Technical Field

[0001] The present application relates to the technical field of scanning electron microscopes, and particularly to a method for evaluating the resolution of a scanning electron microscope, a computer device, and a storage medium. Background Art

[0002] A scanning electron microscope (SEM) is an instrument that uses an electron beam to scan the surface of a sample and forms an image by collecting signals such as secondary electrons and backscattered electrons. In practical applications, an SEM is usually used to scan microscopic samples, and then observe the specific details on the surface of the microscopic samples. Therefore, the resolution of the SEM, that is, the ability to resolve the specific details on the surface of the microscopic samples, is crucial for it. However, the current methods for evaluating the resolution of a scanning electron microscope usually have low accuracy. Summary of the Invention

[0003] The purpose of the embodiments of the present application is to provide a method for evaluating the resolution of a scanning electron microscope, a computer device, and a storage medium to solve the problem of low accuracy in the prior art.

[0004] To solve the above technical problems, the embodiments of the present application provide a method for evaluating the resolution of a scanning electron microscope, which adopts the following technical solutions:

[0005] Obtain a scanned image of a sample by the scanning electron microscope to be evaluated;

[0006] Use an image enhancement model to perform image enhancement processing on the scanned image to obtain an enhanced scanned image, wherein the area containing at least pores in the enhanced scanned image is enhanced to facilitate determining the size of each pore and the pore spacing between pores;

[0007] Evaluate the resolution of the scanning electron microscope to be evaluated through the enhanced scanned image.

[0008] Preferably, the method further includes:

[0009] Obtain an original image dataset, where the original image dataset includes a plurality of different scanning electron microscopes, respectively performing image acquisition on a plurality of different samples to obtain a plurality of original sample images;

[0010] Generate labels corresponding to each original sample image in the original image dataset to generate a plurality of training samples, where the labels include the average size of pores and the number of pores in the corresponding original sample image; the training samples are used to train and obtain the image enhancement model.

[0011] Preferably, generating labels corresponding to each original sample image in the original image dataset specifically includes:

[0012] Take each original sample image in the original image dataset as the current original sample image, and use a target detection algorithm to preliminarily detect the pores in the current original sample image to preliminarily determine the positions of the pores in the current original sample image;

[0013] According to the preliminarily determined positions of the pores in the current original sample image, divide the current original sample image into multiple image regions containing pores, and count the number of pores in each image region;

[0014] For each image region in the current original sample image, use morphological operations and edge detection algorithms to accurately extract the edges of each pore in the image region, and determine the sizes of each pore according to the edges of each pore;

[0015] Calculate the average size of the pores in the current original sample image according to the number of pores in each image region and the sizes of each pore;

[0016] Generate the label corresponding to the current original sample image by using the average size and the number of pores in the current original sample image.

[0017] Preferably, the method further includes:

[0018] Input the original sample images in each training sample into the basic model respectively to obtain the output result of the basic model, and use a double loss function to calculate the average error between the output result and the label in the training sample, and use the average error as negative feedback to adjust the parameters of the basic model.

[0019] Preferably, using a double loss function to calculate the average error between the output result and the label in the training sample specifically includes:

[0020] Use two loss functions to calculate the error between the output result and the label respectively;

[0021] Calculate the average error through the formula L = w1×L1 + w2×L2;

[0022] Where, w1 and w2 are both weights; L1 is the error between the output result and the label calculated by the first loss function; L2 is the error between the output result and the label calculated by the second loss function; L is the calculated average error.

[0023] Preferably, obtain the basic model in the following manner:

[0024] Obtain a convolutional neural network;

[0025] Increase the number of convolutional layers in the convolutional neural network to 2 to 3 times the original number of layers;

[0026] In the convolutional neural network after increasing the number of convolutional layers, adjust the number of convolutional kernels to 1 / 4 to 1 / 2 of the original number of convolutional kernels;

[0027] For the convolutional neural network after adjusting the number of convolutional kernels, delete the pooling layer in the convolutional neural network to obtain the basic model.

[0028] Preferably, evaluate the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanning image, which specifically includes:

[0029] Analyze the enhanced scanning image to determine the pore size of each pore and the pore spacing between pores in the enhanced scanning image;

[0030] Obtain the minimum pore size and the minimum pore spacing from the pore sizes of each pore and the pore spacing between pores in the enhanced scanning image. The minimum pore size and the minimum pore spacing can reflect the limit of the resolution ability of the scanning electron microscope to be evaluated;

[0031] Evaluate the resolution ability of the scanning electron microscope to be evaluated based on the minimum pore size and the minimum pore spacing.

[0032] Preferably, before using the image enhancement model to perform image enhancement processing on the scanning image, it specifically includes:

[0033] Evaluate the image quality of the scanning image to obtain a quality evaluation value of the scanning image;

[0034] Judge whether the quality evaluation value of the image quality is less than or equal to a preset minimum value;

[0035] If so, feedback a prompt message for resubmitting the scanning image to the client application; or,

[0036] If not, input the scanning image into the image enhancement model and use the image enhancement model to perform image enhancement processing on the scanning image.

[0037] To solve the above technical problems, an embodiment of the present application also provides a computer device, which adopts the following technical solutions:

[0038] It includes a memory and a processor. Computer-readable instructions are stored in the memory, and when the processor executes the computer-readable instructions, the steps of the above method are implemented.

[0039] To solve the above technical problems, an embodiment of the present application further provides a computer-readable storage medium, adopting the following technical solution:

[0040] Computer-readable instructions are stored on the computer-readable storage medium, and when the computer-readable instructions are executed by a processor, the steps of the above-described method are implemented.

[0041] Adopting the scanning electron microscope resolution evaluation method provided by the embodiment of the present application, the method includes obtaining a scanned image of a sample by the scanning electron microscope to be evaluated, and then performing image enhancement processing on the scanned image by using an image enhancement model, so as to obtain an enhanced scanned image, and then evaluating the resolution of the scanning electron microscope to be evaluated through the enhanced scanned image. On the one hand, the method uses the sample as the scanning object of the scanning electron microscope to be evaluated, and utilizes the material properties of the mesoporous molecular sieve to improve the accuracy of the evaluation result. On the other hand, since the resolution of the scanning electron microscope to be evaluated is evaluated through the enhanced scanned image, the enhanced scanned image has higher image quality compared with the scanned image before enhancement. Therefore, evaluating the resolution of the scanning electron microscope to be evaluated through the enhanced scanned image can obviously also improve the accuracy of the evaluation result. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the solutions in the present application, the following will briefly introduce the drawings required for the description of the embodiments of the present application. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0043] Figure 1 is an exemplary system architecture diagram to which the present application can be applied;

[0044] Figure 2 is the implementation flowchart of the scanning electron microscope resolution evaluation method provided by the embodiment of the present application;

[0045] Figure 3 is the implementation flowchart of the image enhancement model training method provided by the embodiment of the present application;

[0046] Figure 4 is the structural schematic diagram of the scanning electron microscope resolution evaluation device provided by the embodiment of the present application;

[0047] Figure 5 is the structural schematic diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0048] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs; the terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification and claims of this application or the above drawings are used to distinguish different objects and not to describe a specific order.

[0049] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in connection with the embodiments can be included in at least one embodiment of this application. The phrase appears in various places in the specification and does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment mutually exclusive with other embodiments. Those skilled in the art will explicitly and implicitly understand that the embodiments described herein can be combined with other embodiments.

[0050] To enable those skilled in the technical field to better understand the solution of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings.

[0051] As Figure 1 shown, the system architecture 100 may include a terminal device 101, a network 102, and a server 103. The terminal device 101 may be a laptop computer 1011, a tablet computer 1012, or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links, or fiber optic cables, etc.

[0052] A user may use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various client applications may be installed on the terminal device 101 to facilitate interaction with the server 103 through these client applications.

[0053] The terminal device 101 can be various electronic devices with a display screen and supporting web browsing. In addition to the laptop 1011, tablet computer 1012, or mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV) player, a laptop portable computer, a desktop computer, and so on.

[0054] The server 103 can be a server for evaluating the resolution ability of a scanning electron microscope. For example, based on the scanning image sent by the terminal device 101, it evaluates the resolution ability of the scanning electron microscope that obtained the scanning image.

[0055] It should be noted that the method provided by the embodiments of the present application is generally executed by the server / terminal device, and the corresponding device is generally set in the server / terminal device.

[0056] It should be understood that Figure 1 the numbers of the terminal devices, networks, and servers in

[0057] Continuing to refer to Figure 2 , a flowchart of an embodiment of the method for evaluating the resolution ability of a scanning electron microscope provided by the embodiments of the present application is shown. For ease of understanding, the basic principle of this method can be briefly described here.

[0058] In practical applications, for a scanning electron microscope whose resolution ability needs to be evaluated (hereinafter referred to as the scanning electron microscope to be evaluated), theoretically, any object's surface can be scanned by using the scanning electron microscope to be evaluated to obtain a scanning image, and then the resolution ability can be evaluated through the method of the present application. However, in the present application, in order to improve the accuracy of the evaluation result, it is selected to scan a sample by using the scanning electron microscope to be evaluated, so as to obtain a scanning image.

[0059] Among them, the sample can be a mesoporous molecular sieve sample, or a metal particle sample such as gold particles, silver particles, and tin ball particles. Among them, the mesoporous molecular sieve is a new type of material with a pore diameter between micropores and macropores, having a huge surface area and a three-dimensional pore structure. Its pore diameter is usually between 2 nm and 50 nm, the pore structure is regular and the distribution is narrow. Common mesoporous molecular sieves include SBA-15, MCM-41, SBA-16, etc.

[0060] The samples selected in this application can be, for example, mesoporous molecular sieve samples, that is, mesoporous molecular sieves are used as samples. The mesoporous molecular sieve can be a powdered mesoporous molecular sieve or a mesoporous molecular sieve in other forms. Among them, the reason for selecting the sample is that on the one hand, the pore structure of the mesoporous molecular sieve is regular and clear, which is conducive to observing the imaging effect of the scanning electron microscope; on the other hand, the pore size of the mesoporous molecular sieve is within the resolution range of the scanning electron microscope, and is neither too large nor too small relative to the resolution range of the scanning electron microscope, so it can accurately reflect the resolution ability of the scanning electron microscope; on the third hand, the surface physical and chemical properties of the mesoporous molecular sieve are stable and are not easily affected by the electron beam and change, which is conducive to ensuring the stability of the test results.

[0061] Similarly, the samples selected in this application can be, for example, gold particles, silver particles, and tin ball particles. The reason is that the boundaries of these metal particles are regular and clear, which is conducive to observing the imaging effect of the scanning electron microscope, and particles within a specific size range can be selected, so it can accurately reflect the resolution ability of the scanning electron microscope. In addition, the physical and chemical properties of these metals are also relatively stable and are not easily affected by the electron beam and change, which is conducive to ensuring the stability of the test results. Therefore, these metal particles can also be selected as samples, that is, metal particle samples.

[0062] In addition, after obtaining the scanning image in this application, the resolution ability of the scanning electron microscope to be evaluated is not directly evaluated through this scanning image, but the scanning image is first subjected to image enhancement processing to obtain an enhanced scanning image, and then the resolution ability of the scanning electron microscope to be evaluated is evaluated through the enhanced scanning image.

[0063] In Figure 2 In the method shown, some or all of the steps of this method can be executed by the Figure 1 server 103 or the terminal device 101 shown. For example, here it can be taken as an example that the server 103 executes this method. In this example, the sample can be specifically a mesoporous molecular sieve sample to illustrate this method. This method includes the following steps:

[0064] Step S21: Obtain a scanning image of the sample by the scanning electron microscope to be evaluated.

[0065] In practical applications, the sample can be prepared first. The specific method can be, for example, dispersing the mesoporous molecular sieve powder in a solvent. For example, the mesoporous molecular sieve powder can be dispersed in the solvent by ultrasonic vibration to obtain a dispersion. Among them, the solvent can be deionized water or other types of solvents. After dispersing the mesoporous molecular sieve powder in the solvent, a small amount of the dispersion is sucked and dropped on the sample stage. After the solvent evaporates, the mesoporous molecular sieve powder on the sample stage can be sputter-coated with gold to obtain the sample. Through this sputter-coating treatment, the conductivity of the sample can be improved.

[0066] Of course, after obtaining the sample, the sample can be scanned using the scanning electron microscope to be evaluated to obtain the scanning image.

[0067] Therefore, one implementation manner of step S21 can be to scan the sample using the scanning electron microscope to be evaluated to obtain the scanning image.

[0068] Another implementation manner of step S21 can be to obtain the scanning image sent by the client application. For example, after the user of the client application scans the sample using the scanning electron microscope to be evaluated to obtain the scanning image, the scanning image can be sent to the server through the client application, so that the server can obtain the scanning image and then use the scanning image to evaluate the resolution ability of the scanning electron microscope to be evaluated in the subsequent process.

[0069] It should be further noted that since the server is often connected to multiple terminal devices (the client application may be installed in each terminal device), and different terminal devices may belong to different users, these users can all send scanning images to the server through the client application on their own terminal devices, so that the server can evaluate the resolution ability of the corresponding scanning electron microscope to be evaluated.

[0070] In this case, the following situation may occur, that is, in a short period of time, multiple different users need to evaluate the resolution ability of their respective scanning electron microscopes to be evaluated. Therefore, they send scanning images to the server through the client applications on different terminal devices respectively, resulting in an increase in the computing pressure of the server in a short period of time and even causing the server to be blocked.

[0071] In view of this situation, the present application can pre - construct a pre - storage queue on the server. In this way, for the scanned images sent by the client applications on different terminal devices, according to the order of receipt by the server, these scanned images can be added to the pre - storage queue in sequence. Then, according to the server's own business processing situation and in combination with the order of the scanned images pre - stored in the pre - storage queue, the corresponding scanned images are retrieved from the pre - storage queue, and then the subsequent steps in the method of the present application are executed. By this means, the computing pressure on the server can be reduced. At this time, the specific implementation manner of step S21 can be to retrieve the corresponding scanned images from the pre - storage queue according to the order of the scanned images sent by the client applications on each terminal device in the pre - storage queue.

[0072] Step S22: Use an image enhancement model to perform image enhancement processing on the scanned image to obtain an enhanced scanned image.

[0073] Among them, the image enhancement model can be used to perform image enhancement processing on an image, so that the enhanced image has higher image quality compared with the image before enhancement. For example, on the basis of the image before enhancement, errors, interference caused by other electromagnetic signals, and image distortion can be excluded, etc., to achieve image enhancement processing. Through this enhancement processing, at least the area containing pores in the enhanced scanned image can be enhanced, thus facilitating the determination of the size of each pore and the pore spacing between pores.

[0074] The first implementation manner of step S22 can be to directly input the scanned image into the image enhancement model, and then use the image enhancement model to perform image enhancement processing on the scanned image, so as to obtain the output of the image enhancement model, that is, the enhanced scanned image.

[0075] It should be noted that the present application can pre - train the image enhancement model in the following way. Among them, the image enhancement model can be based on a convolutional neural network (CNN), and then the image enhancement model is obtained through training on the basis of this basic model. The specific training process is as follows:

[0076] Step S221: Generate multiple training samples.

[0077] In practical applications, it is usually necessary to use the original image data set to generate the training samples. Therefore, it is necessary to first obtain the original image data set. Among them, the original image data set includes multiple different scanning electron microscopes, and multiple different scanned images (this image is called the original sample image) obtained by respectively collecting images of multiple different samples.

[0078] After obtaining the original image dataset, the labels corresponding to each original sample image in the original image dataset can be further generated to obtain corresponding training samples, where the labels include the average size and the number of pores in the corresponding original sample image. In practical applications, the labels of the original sample images can be obtained by separately annotating each original sample image in the original image dataset.

[0079] When separately annotating each original sample image, it is necessary to first determine the average size and the number of pores in each original sample image. For example, for a certain original sample image, it is necessary to determine the average size and the number of pores in the original sample image, and then use the average size and the number of pores in the original sample image to generate the label corresponding to the original sample image.

[0080] Among them, the average size and the number of pores in each original sample image can be determined in the following way. Specifically, each original sample image in the original image dataset can be taken as the current original sample image, and a target detection algorithm can be used to preliminarily detect the pores in the current original sample image, so as to preliminarily determine the positions of the pores in the current original sample image, where the target detection algorithm can include algorithms such as YOLO and Faster R-CNN.

[0081] After preliminarily determining the positions of the pores in the current original sample image, the current original sample image can be divided into multiple image regions containing pores according to the preliminarily determined positions of the pores in the current original sample image, and the number of pores in each image region can be counted, where the image segmentation algorithm can be algorithms such as U-Net and Mask R-CNN. For the specific division method, according to the preliminarily determined positions of the pores in the current original sample image, the number of pores in each of the divided image regions can be made roughly the same, so as to avoid too many pores in a few image regions and affect the accuracy.

[0082] After dividing the current original sample image into multiple image regions containing pores, for each image region in the current original sample image, morphological operations (such as erosion, dilation, etc.) and edge detection algorithms (such as Canny edge detection) can be used to accurately extract the edges of each pore in the image region, and the size of each pore can be determined according to the edges of each pore.

[0083] Then, according to the number of pores in each image region and the size of each pore, calculate the average size of the pores in the current original sample image. Furthermore, use the average size and the number of pores in the current original sample image to generate a label corresponding to the current original sample image. In this way, a training sample can be generated from the current original sample image and the label corresponding to the current original sample image.

[0084] Similarly, each original sample image in the original image dataset can be used as the current original sample image respectively, and then multiple training samples can be generated.

[0085] Step S222: Pre-train an image enhancement model using the training samples.

[0086] In the embodiments of the present application, for each training sample, the original sample image in the training sample can be input into the basic model to obtain the output result of the basic model. Then, a double loss function is used to calculate the average error between the output result and the label in the training sample, and this average error is used as negative feedback to adjust the parameters of the basic model. In this way, through continuous multi-round training with multiple training samples, the image enhancement model can finally be generated.

[0087] In the present application, a double loss function is adopted. Its principle is that for the original sample image with a larger average pore size, this double loss function can pay more attention to edge sharpening and contrast enhancement, while for the original sample image with a smaller average pore size, it pays more attention to detail preservation and noise removal. Therefore, this design method of the double loss function can directly affect the training of related parameters such as the convolution kernel size, the number of convolutional layers, and the pooling strategy in the convolutional neural network model.

[0088] Among them, the design method of the double loss function of the present application can be to first use two loss functions to calculate the errors between the output result and the label in the training sample respectively, and then calculate the weighted average of the two as the finally calculated average error. For example, the average error can be calculated by the following formula: L = w1×L1 + w2×L2.

[0089] In this formula, both w1 and w2 are weights. In practical applications, both w1 and w2 are greater than 0; L1 is the error between the output result and the label in the training sample calculated by the first loss function; L2 is the error between the output result and the label in the training sample calculated by the second loss function; L is the calculated average error. Among them, the specific functional forms of the first loss function and the second loss function are not specifically limited here.

[0090] In this design method of the dual loss function, the average error calculated by the dual loss function is used as negative feedback to adjust the parameters of the basic model, so as to perform continuous multiple rounds of training through multiple training samples, and finally generate the image enhancement model.

[0091] It should be noted that the above-mentioned application can use a convolutional neural network as the basic model. However, considering that the application scenario of this application is the original sample images with a relatively small average size of a large number of pores, the recognition of image detail information is very crucial. Therefore, this application can also optimize the convolutional neural network and use the optimized convolutional neural network as the basic model.

[0092] Among them, the convolutional neural network can be optimized in the following ways:

[0093] First, reduce the convolutional kernels of the convolutional neural network (for example, it can be reduced to about 1 / 4 to half of the original convolutional kernels). Since it is necessary to capture the image detail information in the original sample images, it is necessary to reduce the convolutional kernels of the convolutional neural network, so as to adopt relatively small convolutional kernels and reduce the loss of information.

[0094] Second, increase the number of convolutional layers of the convolutional neural network. In this application, it is necessary to increase the number of convolutional layers of the convolutional neural network (for example, it can be increased to 2 to 3 times the original number of layers), so as to extract image features more deeply, including gradually abstracting high-level features in the image while retaining necessary detail information.

[0095] Third, delete the pooling layer in the convolutional neural network. In this application, it is necessary to avoid using the pooling layer to avoid the loss of details caused by excessive downsampling.

[0096] Therefore, in this application, a convolutional neural network can be obtained, and then the convolutional neural network can be optimized in this way, that is, reducing its convolutional kernels, increasing its number of convolutional layers, and deleting its pooling layer, and then using the optimized convolutional neural network as the basic model. For example, in an embodiment, a convolutional neural network can be obtained first, then the number of convolutional layers in the convolutional neural network can be increased to 2 to 3 times the original number of layers, and then the number of convolutional kernels in the convolutional neural network after increasing the number of convolutional layers can be adjusted to 1 / 4 to 1 / 2 of the original number of convolutional kernels, and then for the convolutional neural network after adjusting the number of convolutional kernels, delete the pooling layer in the convolutional neural network to obtain the basic model, and then train the basic model to obtain the image enhancement model.

[0097] Step S23: Evaluate the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanning image.

[0098] After obtaining the enhanced scanned image through the above-mentioned step S22, in this step S23, the resolution ability of the scanning electron microscope to be evaluated can be evaluated through the enhanced scanned image. Since the enhanced scanned image is obtained by enhancing the scanned image, the enhanced scanned image has higher image quality compared to the scanned image before enhancement, including excluding errors caused by the vibration of the scanning electron microscope, interference caused by other electromagnetic signals, and image distortion, etc. on the basis of the scanned image before enhancement. Furthermore, the area containing at least the pores in the enhanced scanned image is enhanced, facilitating the determination of the sizes of each pore and the pore spacing between pores. Therefore, the enhanced scanned image can obviously better reflect the true resolution ability of the scanning electron microscope. In this step S23, evaluating the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanned image is obviously more accurate.

[0099] Among them, for the specific implementation manner of this step S23, for example, by analyzing the enhanced scanned image to determine the pore sizes of each pore and the pore spacing between pores in the enhanced scanned image, and then obtaining the minimum pore size and the minimum pore spacing from the pore sizes of each pore and the pore spacing between pores in the enhanced scanned image. The minimum pore size and the minimum pore spacing can reflect the limit of the resolution ability of the scanning electron microscope to be evaluated. Therefore, based on the minimum pore size and the minimum pore spacing, the resolution ability of the scanning electron microscope to be evaluated is evaluated. Among them, if the minimum pore size and the minimum pore spacing are smaller, it indicates that the resolution ability of the scanning electron microscope to be evaluated is higher; on the contrary, if the minimum pore size and the minimum pore spacing are larger, it indicates that the resolution ability of the scanning electron microscope to be evaluated is lower.

[0100] Adopting the method for evaluating the resolution ability of a scanning electron microscope provided by the embodiment of the present application, this method includes obtaining a scanned image of a sample by the scanning electron microscope to be evaluated, then using an image enhancement model to perform image enhancement processing on the scanned image to obtain an enhanced scanned image, and then evaluating the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanned image. On the one hand, this method uses the sample as the scanning object of the scanning electron microscope to be evaluated, and utilizes the material properties of mesoporous molecular sieves to improve the accuracy of the evaluation result. On the other hand, since the resolution ability of the scanning electron microscope to be evaluated is evaluated through the enhanced scanned image, and the enhanced scanned image has higher image quality compared to the scanned image before enhancement. Therefore, evaluating the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanned image can obviously also improve the accuracy of the evaluation result.

[0101] As mentioned in the above step S22, the first implementation manner of step S22 may be to directly input the scanned image into the image enhancement model, and then use the image enhancement model to perform image enhancement processing on the scanned image, so as to obtain the output of the image enhancement model, that is, the enhanced scanned image.

[0102] It should be further noted that in practical applications, before performing step S22, the image quality of the scanned image can also be evaluated first to obtain the quality evaluation value of the scanned image, and then it is judged whether the quality evaluation value of the image quality is less than or equal to a preset minimum value. At this time, if the quality evaluation value is less than or equal to the preset minimum value, it means that the image quality of the scanned image is too low, and a prompt message for re-submitting the scanned image can be fed back to the client application. If the quality evaluation value is greater than the preset minimum value, it means that the image quality of the scanned image meets the lowest quality requirements, and then step S22 can be executed, that is, the scanned image can be input into the image enhancement model, and then the image enhancement model is used to perform image enhancement processing on the scanned image, so as to obtain the output of the image enhancement model, that is, the enhanced scanned image.

[0103] In this application, the specific manner of evaluating the image quality of the scanned image may be to comprehensively calculate the quality evaluation value of the scanned image by combining the resolution, signal-to-noise ratio, and image distortion degree of the scanned image. Among them, the resolution reflects the minimum detail size that the scanned image can clearly display. In this application, since the scanning electron microscope will record the relevant data during the scanning process after scanning the scanned image, the resolution can be directly obtained from the recorded relevant data; the signal-to-noise ratio refers to the ratio of the signal intensity to the background noise, and the signal-to-noise ratio of the scanned image can be measured by relevant image analysis software; the image distortion degree refers to the deviation between the shape and proportion of the sample in the scanned image and the actual sample, and the image distortion degree can be obtained by comparing the scanned image with the geometric features of the standard sample of mesoporous molecular sieve powder. Therefore, in this application, the resolution (R), signal-to-noise ratio (S / N), and image distortion degree (D) of the scanned image can be obtained through this method.

[0104] After obtaining the resolution, signal-to-noise ratio, and image distortion degree of the scanned image in this way, the evaluation values corresponding to the resolution, signal-to-noise ratio, and image distortion degree can be further calculated, that is, the resolution quality evaluation value, signal-to-noise ratio quality evaluation value, and image distortion degree quality evaluation value. Specifically, the resolution quality evaluation value, signal-to-noise ratio quality evaluation value, and image distortion degree quality evaluation value can be calculated through the following formulas (1) to (3):

[0105]

[0106] In Formulas 1-3, R norm is the calculated resolution quality evaluation value; S / N norm is the calculated signal-to-noise ratio quality evaluation value; D norm is the calculated image distortion degree quality evaluation value; R is the resolution of the scanned image; S / N is the signal-to-noise ratio of the scanned image; D is the image distortion degree of the scanned image; R max is the maximum value of the resolution; R min is the minimum value that the resolution can tolerate; S / N max is the maximum value of the signal-to-noise ratio; S / N min is the minimum value that the signal-to-noise ratio can tolerate; D max is the maximum value that the image distortion degree can tolerate; D min is the minimum value of the image distortion degree.

[0107] After calculating the resolution quality evaluation value, the signal-to-noise ratio quality evaluation value, and the image distortion degree quality evaluation value through the above Formulas 1-3, the average value of the three can be calculated as the quality evaluation value of the scanned image obtained by comprehensive calculation.

[0108] The method of the embodiment of the present application is specifically described by taking a mesoporous molecular sieve sample as an example. In practical applications, based on the same implementation principle, metal particles such as gold particles, silver particles, and tin ball particles can also be used as samples to achieve the invention purpose of the present application. The technical solutions formed in sequence should also be within the protection scope of the present application, and will not be elaborated here one by one.

[0109] Those of ordinary skill in the art can understand that all or part of the processes in the method of the above embodiment can be completed by instructing relevant hardware through computer-readable instructions. The computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM), etc.

[0110] It should be understood that although the steps in the flowchart of the accompanying drawings are shown sequentially in the direction of the arrows, these steps are not necessarily executed sequentially in the order indicated by the arrows. Unless otherwise clearly stated in this document, there is no strict order restriction for the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed alternately or in turn with at least a part of other steps or sub-steps or stages of other steps.

[0111] Based on the same inventive concept as the scanning electron microscope resolution evaluation method provided in the embodiments of the present application, the embodiments of the present application also provide a scanning electron microscope resolution evaluation device. For the content in the embodiments of this device, if there is any unclear point, reference can be made to the corresponding content in the method embodiments. As Figure 4 shown in the specific structural schematic diagram of the device 40, the device 40 includes: an acquisition unit 401, an enhancement processing unit 402, and an evaluation unit 403, where:

[0112] The acquisition unit 401 is configured to acquire a scanned image of a sample by the scanning electron microscope to be evaluated;

[0113] The enhancement processing unit 402 is configured to perform image enhancement processing on the scanned image by using an image enhancement model to obtain an enhanced scanned image, where at least the area containing pores in the enhanced scanned image is enhanced to facilitate determining the sizes of the respective pores and the pore spacing between the pores;

[0114] The evaluation unit 403 is configured to evaluate the resolution of the scanning electron microscope to be evaluated through the enhanced scanned image.

[0115] Since the device 40 adopts the same inventive concept as the scanning electron microscope resolution evaluation method provided in the embodiments of the present application, it can also solve the problems in the prior art, which will not be elaborated here.

[0116] The device 40 may further include a training sample generation unit, configured to acquire an original image data set, where the original image data set includes multiple different scanning electron microscopes, and respectively perform image acquisition on multiple different samples to obtain multiple original sample images; generate labels corresponding to each of the original sample images in the original image data set to generate multiple training samples, where the labels include the average size of the pores and the number of pores in the corresponding original sample image; and the training samples are used to train and obtain the image enhancement model.

[0117] Among them, generating the labels corresponding to the original sample images in the original image dataset may specifically include:

[0118] Taking each original sample image in the original image dataset as the current original sample image, and using a target detection algorithm to preliminarily detect the pores in the current original sample image, so as to preliminarily determine the positions of the pores in the current original sample image;

[0119] According to the preliminarily determined positions of the pores in the current original sample image, dividing the current original sample image into multiple image regions containing pores, and counting the number of pores in each image region;

[0120] For each image region in the current original sample image respectively, using morphological operations and edge detection algorithms to accurately extract the edges of each pore in the image region, and determining the sizes of each pore according to the edges of each pore;

[0121] Calculating the average size of the pores in the current original sample image according to the number of pores in each image region and the sizes of each pore;

[0122] Generating the label corresponding to the current original sample image by using the average size and the number of pores in the current original sample image.

[0123] Among them, the device 40 may further include a model training unit, which is configured to input the original sample images in each training sample into the basic model respectively to obtain the output result of the basic model, and use a double loss function to calculate the average error between the output result and the label in the training sample, and use the average error as negative feedback to adjust the parameters of the basic model.

[0124] Among them, using a double loss function to calculate the average error between the output result and the label in the training sample may specifically include:

[0125] Using two loss functions to calculate the error between the output result and the label respectively;

[0126] Calculating the average error through the formula L = w1×L1 + w2×L2;

[0127] Among them, both w1 and w2 are weights; L1 is the error between the output result and the label calculated by the first loss function; L2 is the error between the output result and the label calculated by the second loss function; L is the calculated average error.

[0128] Among them, the basic model can be obtained in the following way:

[0129] Obtain a convolutional neural network;

[0130] Increase the number of convolutional layers in the convolutional neural network to 2 to 3 times the original number of layers;

[0131] In the convolutional neural network with the increased number of convolutional layers, adjust the number of convolutional kernels to 1 / 4 to 1 / 2 of the original number of convolutional kernels;

[0132] For the convolutional neural network with the adjusted number of convolutional kernels, delete the pooling layer in the convolutional neural network to obtain the basic model.

[0133] Among them, evaluating the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanning image can specifically include:

[0134] Analyze the enhanced scanning image to determine the pore size of each pore and the pore spacing between pores in the enhanced scanning image;

[0135] Obtain the minimum pore size and the minimum pore spacing from the pore size of each pore and the pore spacing between pores in the enhanced scanning image, and the minimum pore size and the minimum pore spacing can reflect the limit of the resolution ability of the scanning electron microscope to be evaluated;

[0136] Evaluate the resolution ability of the scanning electron microscope to be evaluated based on the minimum pore size and the minimum pore spacing.

[0137] Among them, before using the image enhancement model to perform image enhancement processing on the scanning image, it can specifically include:

[0138] Evaluate the image quality of the scanning image to obtain a quality evaluation value of the scanning image;

[0139] Judge whether the quality evaluation value of the image quality is less than or equal to a preset minimum value;

[0140] If so, feedback a prompt message for resubmitting the scanning image to the client application; or,

[0141] If not, input the scanning image into the image enhancement model and use the image enhancement model to perform image enhancement processing on the scanning image.

[0142] Among them, evaluating the image quality of the scanning image specifically includes:

[0143] Determine the resolution, signal-to-noise ratio, and image distortion degree of the scanning image;

[0144] Calculate a resolution quality evaluation value, a signal-to-noise ratio quality evaluation value, and an image distortion quality evaluation value according to the resolution, signal-to-noise ratio, and image distortion degree of the scanned image:

[0145] Calculate the average value of the resolution quality evaluation value, the signal-to-noise ratio quality evaluation value, and the image distortion quality evaluation value as the quality evaluation value of the scanned image.

[0146] To solve the above technical problems, an embodiment of the present application further provides a computer device. For details, please refer to Figure 5 , Figure 5 which is a basic structural block diagram of the computer device according to the embodiment of the present application.

[0147] The computer device 500 includes a memory 510, a processor 520, and a network interface 530 that are communicatively connected to each other through a system bus. It should be noted that only the computer device 500 with components 510-530 is shown in the figure, but it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Among them, those skilled in the art of the present technology can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0148] The computer device can be a computing device such as a desktop computer, a notebook, a palm computer, and a cloud server. The computer device can perform human-computer interaction with the user through a keyboard, a mouse, a remote control, a touchpad, or a voice control device.

[0149] The memory 510 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (such as SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 510 may be an internal storage unit of the computer device 500, such as the hard disk or memory of the computer device 500. In other embodiments, the memory 510 may also be an external storage device of the computer device 500, such as a plug-in hard disk, Smart Media Card (SMC), Secure Digital (SD) card, Flash Card, etc. equipped on the computer device 500. Of course, the memory 510 may also include both the internal storage unit and the external storage device of the computer device 500. In the embodiments of the present application, the memory 510 is generally used to store the operating system and various application software installed on the computer device 500, such as computer-readable instructions of the method provided in the embodiments of the present application. In addition, the memory 510 may also be used to temporarily store various types of data that have been output or will be output.

[0150] In some embodiments, the processor 520 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 520 is generally used to control the overall operation of the computer device 500. In the embodiments of the present application, the processor 520 is used to run the computer-readable instructions stored in the memory 510 or process data, such as running the computer-readable instructions of the method provided in the embodiments of the present application.

[0151] The network interface 530 may include a wireless network interface or a wired network interface, and the network interface 530 is generally used to establish a communication connection between the computer device 500 and other electronic devices.

[0152] The present application also provides another implementation manner, that is, to provide a computer-readable storage medium storing computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to execute the steps of the method as described above.

[0153] Through the description of the above embodiments, those skilled in the art can clearly understand that the above-described example methods can be implemented by means of software plus a necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation. Based on such an understanding, the technical solution of the present application, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions for causing a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of the present application.

[0154] Obviously, the above-described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The accompanying drawings show the preferred embodiments of the present application, but do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosed content of the present application more thorough and comprehensive. Although the present application has been described in detail with reference to the foregoing embodiments, for those skilled in the art, they can still modify the technical solutions described in the foregoing specific embodiments, or perform equivalent replacements on some of the technical features. Any equivalent structure directly or indirectly using the content of the specification and drawings of the present application in other related technical fields is equally within the scope of the patent protection of the present application.

Claims

1. A method for evaluating the resolution ability of a scanning electron microscope, characterized in that Including: Obtain a scanned image of a sample by the scanning electron microscope to be evaluated; Use an image enhancement model to perform image enhancement processing on the scanned image to obtain an enhanced scanned image. Among them, the area containing at least the pores in the enhanced scanned image is enhanced to facilitate determining the size of each pore and the pore spacing between pores; Evaluate the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanned image.

2. The method according to claim 1, wherein The method further includes: Obtain an original image dataset, where the original image dataset includes multiple different scanning electron microscopes that respectively collect images of multiple different samples to obtain multiple original sample images; Generate labels corresponding to each original sample image in the original image dataset to generate multiple training samples. Among them, the label includes the average size of the pores and the number of pores in the corresponding original sample image; the training samples are used to train and obtain the image enhancement model.

3. The method according to claim 2, wherein Generating the labels corresponding to each original sample image in the original image dataset specifically includes: Respectively take each original sample image in the original image dataset as the current original sample image, and use an object detection algorithm to preliminarily detect the pores in the current original sample image to preliminarily determine the positions of the pores in the current original sample image; According to the preliminarily determined positions of the pores in the current original sample image, divide the current original sample image into multiple image regions containing pores, and count the number of pores in each image region; For each image region in the current original sample image, use morphological operations and edge detection algorithms to accurately extract the edges of each pore in the image region, and determine the size of each pore according to the edges of each pore; Calculate the average size of the pores in the current original sample image according to the number of pores in each image region and the size of each pore; Generate a label corresponding to the current original sample image using the average size and the number of pores in the current original sample image.

4. The method according to claim 2, wherein The method further includes: Respectively input the original sample images in each training sample into the basic model to obtain the output result of the basic model, and use a double loss function to calculate the average error between the output result and the label in the training sample, and use the average error as negative feedback to adjust the parameters of the basic model.

5. The method according to claim 4, characterized in that, Using a double loss function to calculate the average error between the output result and the label in the training sample specifically includes: Use two loss functions to calculate the error between the output result and the label respectively; Calculate the average error through the formula L = w1×L1 + w2×L2; Among them, w1 and w2 are both weights; L1 is the error between the output result and the label calculated by the first loss function; L2 is the error between the output result and the label calculated by the second loss function; L is the calculated average error.

6. The method according to claim 4, wherein Obtain the basic model in the following way: Obtain a convolutional neural network; Increase the number of convolutional layers in the convolutional neural network to 2 - 3 times the original number of layers; In the convolutional neural network after increasing the number of convolutional layers, adjust the number of convolutional kernels to 1 / 4 to 1 / 2 of the original number of convolutional kernels; For the convolutional neural network after adjusting the number of convolutional kernels, delete the pooling layer in the convolutional neural network to obtain the basic model.

7. The method according to claim 1, wherein Evaluate the resolution ability of the scanning electron microscope to be evaluated through the enhanced scanning image, specifically including: Analyze the enhanced scanning image to determine the pore size and pore spacing between pores in the enhanced scanning image; Obtain the minimum pore size and the minimum pore spacing from the pore sizes of the pores in the enhanced scanning image and the pore spacing between the pores, and the minimum pore size and the minimum pore spacing can reflect the limit of the resolution ability of the scanning electron microscope to be evaluated; Evaluate the resolution ability of the scanning electron microscope to be evaluated based on the minimum pore size and the minimum pore spacing.

8. The method according to claim 1, wherein Before using the image enhancement model to perform image enhancement processing on the scanning image, specifically including: Evaluate the image quality of the scanning image to obtain a quality evaluation value of the scanning image; Judge whether the quality evaluation value of the image quality is less than or equal to a preset minimum value; If so, feedback a prompt message for resubmitting the scanning image to the client application; or, If not, input the scanning image into the image enhancement model and use the image enhancement model to perform image enhancement processing on the scanning image.

9. A computer device, comprising a memory and a processor, characterized in that, The memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images