Ammeter large model training method and device, medium and equipment

Abstract

The invention provides an ammeter large model training method and device, a medium and equipment. The method comprises the following steps: acquiring a plurality of electric meter images under various environmental conditions; performing feature extraction on each electric meter image to obtain image features of the electric meter image; marking the image features of the electric meter image according to the reading in each electric meter image to obtain a sample; dividing a plurality of samples corresponding to each electric meter image into a training set and a test set; performing model training by using the training set to obtain a convergent electric meter large model; and determining a performance index of the electric meter large model by using the test set, and deploying a model file of the electric meter large model at a target node after the performance index meets a preset requirement. According to the embodiment of the invention, automatic meter reading of the electric meter can be realized through the large electric meter model obtained through training, the reading of the electric meter can be accurately identified, the error rate of manual meter reading is reduced, the labor cost and the time cost are reduced, and the efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of model training, and in particular to a method, device, medium and equipment for training a large model of an electric meter. Background Art

[0002] At present, electricity meters are usually read manually, which not only wastes manpower, but also has a high time cost and the possibility of misreading. Therefore, it is necessary to make improvements. Summary of the invention

[0003] In response to at least one of the above technical problems, an embodiment of the present invention provides a method, device, medium, and equipment for training a large model of an electric meter.

[0004] According to the first aspect, the electric meter large model training method provided by the embodiment of the present invention includes:

[0005] Acquire multiple images of electricity meters under various environmental conditions;

[0006] Perform feature extraction on each electric meter image to obtain image features of the electric meter image;

[0007] Annotate the image features of each meter image according to the readings in the meter image to obtain a sample;

[0008] Divide multiple samples corresponding to each electric meter image into a training set and a test set;

[0009] The training set is used to perform model training to obtain a converged large model of the electric meter; wherein the large model of the electric meter is used to identify the degree of the electric meter;

[0010] The test set is used to determine the performance index of the large model of the electric meter, and after the performance index reaches the preset requirement, the model file of the large model of the electric meter is deployed on the target node.

[0011] In one embodiment, before extracting features from each electric meter image, the method further includes:

[0012] Preprocessing the plurality of electric meter images; wherein the preprocessing includes cleaning;

[0013] Data enhancement is performed on the preprocessed electric meter image to obtain an enhanced electric meter image; wherein the enhancement method includes at least one of rotation, scaling and color conversion.

[0014] In one embodiment, the use of the training set for model training includes: in the process of using the training set for model training, adjusting hyperparameters by random search, accelerating the convergence rate of the model by using an adaptive moment estimation optimization algorithm, using a cross entropy loss function to determine whether the model has converged, and using regularization technology to improve the generalization ability of the model.

[0015] In one embodiment, the use of the training set to perform model training includes: using the training set to perform model training on a preset model framework; wherein the preset model framework includes a residual network, a convolutional layer, a pooling layer, a fully connected layer and an activation function.

[0016] In one embodiment, deploying the model file of the large model of the electric meter on the target node includes: converting the model file of the large model of the electric meter into a preset format suitable for the deployment environment, and sending the model file in the preset format to the target node, so that the target node deploys the large model of the electric meter.

[0017] In one embodiment, the method further comprises:

[0018] After the large electric meter model is deployed and applied, newly added electric meter images are regularly acquired, and the newly added electric meter images are annotated using real electric meter readings to obtain newly added samples;

[0019] The newly added samples are used to perform incremental training on the large model of the electric meter, and a model file of the large model of the electric meter obtained after the incremental training is sent to the target node, so that the target node is updated.

[0020] In one embodiment, using the test set to determine the performance index of the electric meter large model includes:

[0021] Determine the meter reading output by the large meter model using the test set;

[0022] The performance indicator is calculated according to the difference between the meter reading output by the meter model and the marked real meter reading.

[0023] According to the second aspect, the electric meter large model training device provided by the embodiment of the present invention includes:

[0024] An image acquisition module, used to acquire multiple images of electric meters under various environmental conditions;

[0025] A feature extraction module is used to extract features from each electric meter image to obtain image features of the electric meter image;

[0026] A feature annotation module is used to annotate the image features of each meter image according to the readings in the meter image to obtain a sample;

[0027] A sample division module, used to divide multiple samples corresponding to each electric meter image into a training set and a test set;

[0028] A model training module, used to perform model training using the training set to obtain a converged large model of the electric meter; wherein the large model of the electric meter is used to identify the degree of the electric meter;

[0029] The model deployment module is used to use the test set to determine the performance index of the large model of the electric meter, and deploy the model file of the large model of the electric meter on the target node after the performance index reaches the preset requirement.

[0030] In one embodiment, before extracting features from each electric meter image, the feature extraction module is further used to: preprocess the multiple electric meter images; wherein the preprocessing includes cleaning; and perform data enhancement on the preprocessed electric meter image to obtain an enhanced electric meter image; wherein the enhancement method includes at least one of rotation, scaling, and color transformation.

[0031] In one embodiment, the model training module is specifically used to: in the process of model training using the training set, adjust hyperparameters by random search, accelerate the convergence rate of the model by using an adaptive moment estimation optimization algorithm, use a cross entropy loss function to determine whether the model has converged, and use regularization technology to improve the generalization ability of the model.

[0032] In one embodiment, the model training module is specifically used to: use the training set to perform model training on a preset model framework; wherein the preset model framework includes a residual network, a convolutional layer, a pooling layer, a fully connected layer and an activation function.

[0033] In one embodiment, the model deployment module is specifically used to: convert the model file of the large electric meter model into a preset format suitable for the deployment environment, and send the model file in the preset format to the target node so that the target node deploys the large electric meter model.

[0034] In one embodiment, the device may further include:

[0035] The incremental update module is used to periodically acquire new meter images after the meter big model is deployed and applied, and to annotate the new meter images with real meter readings to obtain new samples; to perform incremental training on the meter big model with the new samples, and to send the model file of the meter big model obtained after the incremental training to the target node so that the target node is updated.

[0036] In one embodiment, the model deployment module is specifically used to: determine the meter reading output by the large meter model using the test set; and calculate the performance indicator based on the gap between the meter reading output by the large meter model and the marked actual meter reading.

[0037] According to a third aspect, an embodiment of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed in a computer, enables the computer to execute the method provided in the first aspect.

[0038] According to a fourth aspect, an embodiment of the present invention provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, the method provided in the first aspect is implemented.

[0039] The method, device, medium, and equipment for training a large model of an electric meter provided in an embodiment of the present invention obtain multiple images of electric meter under multiple environmental conditions, then perform feature extraction on each electric meter image to obtain image features of the electric meter image, then annotate the image features of the electric meter image to obtain samples, divide the multiple samples corresponding to each electric meter image into a training set and a test set, use the training set to perform model training to obtain a converged large model of the electric meter, use the test set to determine the performance index of the large model of the electric meter, and deploy the model file of the large model of the electric meter to the target node after the performance index reaches the preset requirements. It can be seen that the large model of the electric meter obtained by training in the embodiment of the present invention can realize automatic meter reading of the electric meter. This automatic method can accurately identify and read the readings of the electric meter, reduce the error rate of manual meter reading, and reduce labor costs and time costs, thereby improving efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of a flow chart of a method for training a large model of an electric meter in one embodiment of the present invention;

[0041] Figure 2 It is a structural block diagram of an electric meter large model training device in one embodiment of the present invention. DETAILED DESCRIPTION

[0042] In a first aspect, an embodiment of the present invention provides a method for training a large model of an electric meter. Figure 1 The method comprises the following steps S110 to S160:

[0043] S110, acquiring multiple electricity meter images under various environmental conditions;

[0044] Among them, a variety of environmental conditions, for example, under different lighting, angles, etc. By diversifying the shooting conditions, the generalization ability of the model can be improved.

[0045] In one embodiment, before extracting features from each electricity meter image in S120, the method may further include:

[0046] Preprocessing the plurality of electric meter images; wherein the preprocessing includes cleaning;

[0047] Data enhancement is performed on the preprocessed electric meter image to obtain an enhanced electric meter image; wherein the enhancement method includes at least one of rotation, scaling and color conversion.

[0048] That is, the meter images are first preprocessed by cleaning, which can remove poor quality images, such as blurred or irrelevant samples, to ensure the high quality of the training data. Then data enhancement is performed, for example, the meter images are expanded by rotation, scaling, color conversion, etc., which can increase the diversity of samples and reduce the risk of overfitting.

[0049] S120, extracting features from each electric meter image to obtain image features of the electric meter image;

[0050] It is understandable that after feature extraction of each electric meter image, the feature vector obtained is used as the image feature of the electric meter image, and the corresponding electric meter image is represented by the image feature.

[0051] S130, annotating the image features of each electric meter image according to the readings in the electric meter image to obtain a sample;

[0052] It is understandable that the reading used for image feature annotation of an electric meter image is the actual reading of the electric meter image.

[0053] S140, dividing the multiple samples corresponding to each electric meter image into a training set and a test set;

[0054] For example, 80% of the samples form the training set and the remaining 20% ​​of the samples form the test set.

[0055] S150, using the training set to perform model training to obtain a converged large model of the electric meter; wherein the large model of the electric meter is used to identify the degree of the electric meter;

[0056] In one embodiment, the use of the training set to perform model training may include: using the training set to perform model training on a preset model framework; wherein the preset model framework includes a residual network, a convolutional layer, a pooling layer, a fully connected layer and an activation function.

[0057] Among them, in the process of determining the preset model framework, the basic network is first selected. Here, the residual network ResNet is used as the basic network, and the image features of the meter image are processed through the residual network; then, on the basis of the basic network, convolutional layers and pooling layers are added, and the image data obtained after the residual network processing is extracted through the convolutional layers and pooling layers to extract richer features. Finally, a fully connected layer and an activation function are added to the last part of the architecture, for example, a softmax activation function, to output the meter reading.

[0058] It can be seen that by training the model based on the preset model framework, a converged large model of the electricity meter is obtained.

[0059] In one embodiment, the performing model training using the training set may include:

[0060] In the process of model training using the training set, hyperparameters are adjusted through random search, the adaptive moment estimation optimization algorithm is used to accelerate the convergence rate of the model, the cross entropy loss function is used to determine whether the model has converged, and regularization technology is used to improve the generalization ability of the model.

[0061] In other words, the learning rate, batch size and other hyperparameters can be adjusted through random search to find the optimal hyperparameter configuration. The adaptive moment estimation optimization algorithm, namely the Adam optimization algorithm, can accelerate the convergence of the model and improve performance. The cross entropy loss function can be used to determine whether the model has converged. For example, if the loss value calculated by the cross entropy loss function is less than a certain level, it is considered to have converged. Using L1 / L2 regularization technology, overfitting can be reduced and the generalization ability of the model can be improved.

[0062] S160. Determine the performance index of the large electric meter model by using the test set, and deploy the model file of the large electric meter model on the target node after the performance index reaches the preset requirement.

[0063] In one embodiment, determining the performance index of the electric meter large model by using the test set in S160 may include:

[0064] Determine the meter reading output by the large meter model using the test set;

[0065] The performance indicator is calculated according to the difference between the meter reading output by the meter model and the marked real meter reading.

[0066] That is to say, by inputting the image features in the test set into the large model of the electricity meter, we can obtain the reading recognized by the large model of the electricity meter, and then determine the difference between the reading and the real reading marked in the image features, and then calculate the performance index based on the difference between the two. The performance index can reflect the performance of the model.

[0067] It is understandable that if the performance indicators show that the model performance meets certain requirements, the deployment operation will be carried out. Otherwise, the model will be adjusted until the model meets the requirements and then the deployment operation will be carried out.

[0068] In one embodiment, deploying the model file of the large electric meter model on the target node may include: converting the model file of the large electric meter model into a preset format suitable for the deployment environment, and sending the model file in the preset format to the target node, so that the target node deploys the large electric meter model.

[0069] The preset format can be TensorFlow Lite or ONNX.

[0070] It is understandable that converting the model file into a preset format and further compressing the file can reduce the inference delay.

[0071] In one embodiment, the method may further include:

[0072] After the large electric meter model is deployed and applied, newly added electric meter images are regularly acquired, and the newly added electric meter images are annotated using real electric meter readings to obtain newly added samples;

[0073] The newly added samples are used to perform incremental training on the large model of the electric meter, and a model file of the large model of the electric meter obtained after the incremental training is sent to the target node, so that the target node is updated.

[0074] It can be seen that by regularly collecting new electricity meter images, new samples, namely new samples, are formed, and then the model is incrementally trained using the new samples. After training, the model on the target node is updated, so that the model can continuously adapt to environmental changes.

[0075] Of course, the model’s functionality and performance can be continuously updated based on user feedback and technological developments.

[0076] In actual scenarios, an intuitive user interface can also be designed to display model recognition results.

[0077] Among them, large model intelligent recognition technology is an important technology in the current field of artificial intelligence. It combines multiple advanced technologies such as deep learning, machine learning, and big data processing to achieve intelligent recognition of various types of data such as images, text, and voice.

[0078] The following is a detailed summary of existing technologies for large model intelligent recognition:

[0079] Deep learning is one of the core technologies of large-model intelligent recognition. By building a deep neural network, the deep learning model can automatically extract the features of the data and learn and optimize in massive data. In large-model intelligent recognition, deep learning technology is widely used in image recognition, speech recognition, natural language processing and other fields.

[0080] Convolutional Neural Network (CNN): CNN is a deep learning model specifically designed for processing image data. It can effectively extract hierarchical features from the original image through local perception and hierarchical network structure. CNN has a wide range of applications in image recognition, face recognition, autonomous driving and other fields.

[0081] Recurrent Neural Network (RNN): RNN is a neural network model for processing sequence data. It retains historical information through memory units, thereby effectively modeling sequence data. RNN has a wide range of applications in the field of natural language processing, such as speech recognition, machine translation, and text generation.

[0082] Transformer architecture: Transformer is a deep learning model based on the self-attention mechanism, proposed by Google. It captures the characteristics of input data through multi-layer self-attention mechanism and position encoding, and has achieved excellent performance in the field of natural language processing. Transformer has become one of the basic architectures in the field of modern natural language processing.

[0083] The embodiments of the present invention have the following beneficial effects:

[0084] (1) Through the training of deep learning models, the meter readings can be accurately identified and read, reducing the error rate of manual meter reading. This automated process improves accuracy, thereby providing users with reliable energy consumption information and reducing labor costs and time costs. Compared with traditional manual meter reading methods, the automatic recognition method based on large models can significantly improve the efficiency of meter reading. It can process a large amount of image data in a very short time and achieve rapid recognition, which is particularly important for large-scale power grid management.

[0085] (2) In the long run, although it may be necessary to invest in technology development and equipment upgrades in the early stages, this approach can reduce reliance on human resources and reduce long-term operating costs. In addition, reducing human errors also means reducing additional costs caused by errors. The embodiments of the present invention can achieve real-time monitoring of meter readings and provide instant data feedback to power suppliers, which helps to better manage grid loads, optimize energy distribution, and even issue timely alarms when abnormal consumption is detected.

[0086] (3) With the development of smart grid technology, this large model-based meter identification method is combined with it to further improve the intelligence level of the power grid. It provides technical support for the realization of functions such as remote meter reading and dynamic electricity price adjustment. For consumers, this method can provide more accurate and timely electricity consumption information, helping them to better understand their electricity consumption habits and take energy-saving measures. At the same time, it reduces the inconvenience of door-to-door meter reading and improves user satisfaction. By optimizing energy use and management, it helps to reduce energy waste and support environmental protection and sustainable development goals. The deep learning model has good adaptability and can adapt to new meter types and different environmental conditions through continuous learning to maintain long-term effective operation.

[0087] In a second aspect, an embodiment of the present invention provides a large-scale model training device for an electric meter, see Figure 2 , the device 100 comprises:

[0088] An image acquisition module 110, for acquiring a plurality of electric meter images under a variety of environmental conditions;

[0089] A feature extraction module 120 is used to extract features from each electric meter image to obtain image features of the electric meter image;

[0090] A feature annotation module 130 is used to annotate the image features of each meter image according to the readings in the meter image to obtain a sample;

[0091] A sample division module 140 is used to divide a plurality of samples corresponding to each electric meter image into a training set and a test set;

[0092] The model training module 150 is used to perform model training using the training set to obtain a converged large model of the electric meter; wherein the large model of the electric meter is used to identify the degree of the electric meter;

[0093] The model deployment module 160 is used to determine the performance index of the large model of the electric meter by using the test set, and deploy the model file of the large model of the electric meter on the target node after the performance index reaches the preset requirement.

[0094] In one embodiment, before extracting features from each electric meter image, the feature extraction module is further used to: preprocess the multiple electric meter images; wherein the preprocessing includes cleaning; and perform data enhancement on the preprocessed electric meter image to obtain an enhanced electric meter image; wherein the enhancement method includes at least one of rotation, scaling, and color transformation.

[0095] In one embodiment, the model training module is specifically used to: in the process of model training using the training set, adjust hyperparameters by random search, accelerate the convergence rate of the model by using an adaptive moment estimation optimization algorithm, use a cross entropy loss function to determine whether the model has converged, and use regularization technology to improve the generalization ability of the model.

[0096] In one embodiment, the model training module is specifically used to: use the training set to perform model training on a preset model framework; wherein the preset model framework includes a residual network, a convolutional layer, a pooling layer, a fully connected layer and an activation function.

[0097] In one embodiment, the model deployment module is specifically used to: convert the model file of the large electric meter model into a preset format suitable for the deployment environment, and send the model file in the preset format to the target node so that the target node deploys the large electric meter model.

[0098] In one embodiment, the device may further include:

[0099] The incremental update module is used to periodically acquire new meter images after the meter big model is deployed and applied, and to annotate the new meter images with real meter readings to obtain new samples; to perform incremental training on the meter big model with the new samples, and to send the model file of the meter big model obtained after the incremental training to the target node so that the target node is updated.

[0100] In one embodiment, the model deployment module is specifically used to: determine the meter reading output by the large meter model using the test set; and calculate the performance indicator based on the gap between the meter reading output by the large meter model and the marked actual meter reading.

[0101] It is understandable that the explanation, specific implementation, beneficial effects, examples, etc. of the relevant contents in the device provided in the embodiment of the present invention can be found in the corresponding parts of the method provided in the first aspect, and will not be repeated here.

[0102] In a third aspect, an embodiment of the present invention provides a computer-readable medium, on which computer instructions are stored. When the computer instructions are executed by a processor, the processor executes the method provided in the first aspect.

[0103] Specifically, a system or device equipped with a storage medium can be provided, on which software program code that implements the functions of any of the above-mentioned embodiments is stored, and a computer (or CPU or MPU) of the system or device can be enabled to read and execute the program code stored in the storage medium.

[0104] In this case, the program code itself read from the storage medium can realize the function of any one of the above-mentioned embodiments, and thus the program code and the storage medium storing the program code constitute a part of the present invention.

[0105] The storage medium embodiments for providing the program code include a floppy disk, a hard disk, a magneto-optical disk, an optical disk (such as CD-ROM, CD-R, CD-RW, DVD-ROM, DVD-RAM, DVD-RW, DVD+RW), a magnetic tape, a non-volatile memory card, and a ROM. Alternatively, the program code can be downloaded from a server computer by a communication network.

[0106] In addition, it should be clear that the functions of any of the above embodiments can be implemented not only by executing the program code read by the computer, but also by enabling an operating system operating on the computer to complete part or all of the actual operations based on instructions from the program code.

[0107] In addition, it can be understood that the program code read from the storage medium is written to a memory provided in an expansion board inserted into the computer or to a memory provided in an expansion module connected to the computer, and then based on the instructions of the program code, a CPU installed on the expansion board or expansion module is enabled to perform part or all of the actual operations, thereby realizing the functions of any of the above-mentioned embodiments.

[0108] It is understandable that the explanation, specific implementation methods, beneficial effects, examples, etc. of the relevant contents in the computer-readable medium provided in the embodiment of the present invention can be found in the corresponding parts of the method provided in the first aspect, and will not be repeated here.

[0109] In a fourth aspect, an embodiment of the present specification provides a computing device, including a memory and a processor, wherein the memory stores executable code, and when the processor executes the executable code, it implements the method in any embodiment of the specification.

[0110] It is understandable that the explanation, specific implementation, beneficial effects, examples, etc. of the relevant contents in the computing device provided in the embodiment of the present invention can be found in the corresponding parts of the method provided in the first aspect, and will not be repeated here.

[0111] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0112] Those skilled in the art should be aware that in one or more of the above examples, the functions described in the present invention can be implemented by hardware, software, widgets, or any combination thereof. When implemented by software, these functions can be stored in a computer-readable medium or transmitted as one or more instructions or codes on a computer-readable medium.

[0113] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made on the basis of the technical solution of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for training a large model of an electric meter, characterized in that: include: Acquire multiple images of electricity meters under various environmental conditions; Perform feature extraction on each electric meter image to obtain image features of the electric meter image; Annotate the image features of each meter image according to the readings in the meter image to obtain a sample; Divide multiple samples corresponding to each electric meter image into a training set and a test set; The training set is used to perform model training to obtain a converged large model of the electric meter; wherein the large model of the electric meter is used to identify the degree of the electric meter; The test set is used to determine the performance index of the large model of the electric meter, and after the performance index reaches the preset requirement, the model file of the large model of the electric meter is deployed on the target node.

2. The method according to claim 1 is characterized in that Before extracting features from each electric meter image, the method further includes: Preprocessing the plurality of electric meter images; wherein the preprocessing includes cleaning; Data enhancement is performed on the preprocessed electric meter image to obtain an enhanced electric meter image; wherein the enhancement method includes at least one of rotation, scaling and color conversion.

3. The method according to claim 1 is characterized in that The using the training set to perform model training includes: In the process of model training using the training set, hyperparameters are adjusted through random search, the adaptive moment estimation optimization algorithm is used to accelerate the convergence rate of the model, the cross entropy loss function is used to determine whether the model has converged, and regularization technology is used to improve the generalization ability of the model.

4. The method according to claim 1 is characterized in that The using the training set to perform model training includes: The training set is used to perform model training on a preset model framework, wherein the preset model framework includes a residual network, a convolutional layer, a pooling layer, a fully connected layer and an activation function.

5. The method according to claim 1 is characterized in that The step of deploying the model file of the electric meter large model on the target node includes: The model file of the large electric meter model is converted into a preset format suitable for a deployment environment, and the model file in the preset format is sent to the target node, so that the target node deploys the large electric meter model.

6. The method according to claim 1 is characterized in that: Also includes: After the large electric meter model is deployed and applied, newly added electric meter images are regularly acquired, and the newly added electric meter images are annotated using real electric meter readings to obtain newly added samples; The newly added samples are used to perform incremental training on the large model of the electric meter, and a model file of the large model of the electric meter obtained after the incremental training is sent to the target node, so that the target node is updated.

7. The method according to claim 1 is characterized in that: The step of using the test set to determine the performance index of the electric meter large model includes: Determine the meter reading output by the large meter model using the test set; The performance indicator is calculated according to the difference between the meter reading output by the meter model and the marked real meter reading.

8. A large-scale model training device for electric meters, characterized in that: include: An image acquisition module, used to acquire multiple images of electric meters under various environmental conditions; A feature extraction module is used to extract features from each electric meter image to obtain image features of the electric meter image; A feature annotation module is used to annotate the image features of each meter image according to the readings in the meter image to obtain a sample; A sample division module, used to divide multiple samples corresponding to each electric meter image into a training set and a test set; A model training module, used to perform model training using the training set to obtain a converged large model of the electric meter; wherein the large model of the electric meter is used to identify the degree of the electric meter; The model deployment module is used to use the test set to determine the performance index of the large model of the electric meter, and deploy the model file of the large model of the electric meter on the target node after the performance index reaches the preset requirement.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed in a computer, the computer is caused to execute the method according to any one of claims 1 to 7.

10. A computing device, characterized in that The method comprises a memory and a processor, wherein the memory stores executable codes, and when the processor executes the executable codes, the method according to any one of claims 1 to 7 is implemented.

Images