A digital twin method for ultraviolet image data stream of line insulators

Abstract

The present application provides a digital twin method for ultraviolet image data stream of line insulators, which relates to the field of graphic image processing technology, and includes: obtaining observation data of sample insulators under different environments; based on the observation data, performing digital twin modeling on the sample insulators under each type of environment to form at least two first digital twins; preprocessing the first digital twins to remove outliers in the first digital twins to obtain a second digital twin; using the second digital twins to train an initial neural network model to obtain a target neural network model; inputting the insulator twin to be tested into the target neural network model to generate an evaluation result.

Description

Technical Field

[0001] The present invention relates to the field of graphics and image processing technology, and in particular to a digital twin method for ultraviolet image data stream of a line insulator. Background Art

[0002] The drone system has the advantages of strong mobility, agility and safety. It is equipped with ultraviolet optical detection equipment to complete the task of photographing the corona discharge of line insulators. The data of ultraviolet optical detection equipment can be used to build a digital twin of line insulator detection. Through multiple twin comparisons, the operating status of the line insulator can be obtained to avoid damage to the transmission line due to insulator failure. However, the current ultraviolet detection of line insulators often only involves the digital twin of line insulators in a single scenario, and it is difficult to make an accurate judgment on the insulator status in complex scenarios. Summary of the invention

[0003] The purpose of the present invention is to solve the problem that the current ultraviolet detection of line insulators cannot adapt to complex scenes. The present invention provides a digital twin method of ultraviolet image data stream of line insulators.

[0004] The technical solution of the embodiment of the present application is implemented as follows:

[0005] A first aspect of an embodiment of the present application provides a digital twin method of ultraviolet image data stream of a line insulator, comprising:

[0006] Obtain observation data of sample insulators under different environments;

[0007] Based on the observation data, digital twin modeling is performed on the sample insulators in each type of environment to form at least two first digital twins;

[0008] Preprocessing the first digital twin to remove outliers in the first digital twin to obtain a second digital twin;

[0009] Using the second digital twin to train the initial neural network model to obtain a target neural network model;

[0010] The insulator twin to be tested is input into the target neural network model to generate an evaluation result.

[0011] Optionally, preprocessing the first digital twin to remove outliers in the first digital twin to obtain the second digital twin includes:

[0012] Obtaining a standard deviation between each data point in the first digital twin and the corresponding mean;

[0013] Determine data points whose standard deviation is greater than a preset threshold as outliers in the first digital twin;

[0014] The outliers are removed to obtain a second digital twin.

[0015] Optionally, preprocessing the first digital twin to remove outliers in the first digital twin to obtain the second digital twin includes:

[0016] The first digital twin is preprocessed by using a ROR filter to remove outliers in the first digital twin to obtain a second digital twin.

[0017] Optionally, before inputting the insulator twin to be tested into the target neural network model to generate an evaluation result, the method further includes:

[0018] Preprocessing is performed on the insulator twins to be tested, and outliers in the insulator twins to be tested are eliminated.

[0019] Optionally, the initial neural network model is a Grid-GCN model.

[0020] Optionally, the using the second digital twin to train the initial neural network model to obtain a target neural network model includes:

[0021] The initial neural network model is trained using the second digital twin and the corresponding real insulator model to obtain a target neural network model.

[0022] Optionally, a cross entropy loss function is used as a training loss function of the initial neural network model, and the training loss function satisfies:

[0023] ;

[0024] in, is the value of the i-th data point in the real insulator model, is the probability of the i-th category predicted by the target neural network model.

[0025] A second aspect of an embodiment of the present application provides a line insulator ultraviolet image data stream digital twin device, including: an acquisition module, a creation module, a preprocessing module, a training module, and a generation module;

[0026] The acquisition module is configured to acquire observation data of sample insulators under different environments;

[0027] The creation module is configured to perform digital twin modeling on the sample insulators in each type of environment based on the observation data to form at least two first digital twins;

[0028] The preprocessing module is configured to preprocess the first digital twin, remove outliers in the first digital twin, and obtain a second digital twin;

[0029] The training module is configured to train the initial neural network model using the second digital twin to obtain a target neural network model;

[0030] The generation module is configured to input the insulator twin to be tested into the target neural network model to generate an evaluation result.

[0031] A third aspect of an embodiment of the present application provides an electronic device, comprising a processor and a memory; the memory stores a computer program, wherein the computer program, when executed by the processor, implements the line insulator ultraviolet image data stream digital twin method described in the first aspect.

[0032] A fourth aspect of an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the first aspect.

[0033] Compared with the prior art, the technical solution provided by this application has the following beneficial effects:

[0034] The present invention provides a digital twin method, device and equipment for ultraviolet image data stream of line insulators. By obtaining observation data of sample insulators in different environments, digital twin modeling is performed on the sample insulators in each type of environment to form at least two first digital twins. After removing outliers in the first digital twins, the initial neural network model is trained using the digital twins to obtain a target neural network model. The target neural network model is used to judge the situation of the twin of the insulator to be tested, which fully considers the situation of the insulator in complex scenarios and facilitates accurate acquisition of the operating status of the line insulator. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A schematic diagram of a process flow of a line insulator ultraviolet image data stream digital twin method provided in an embodiment of the present application;

[0036] Figure 2 A schematic diagram of the structure of a line insulator ultraviolet image data stream digital twin device provided in an embodiment of the present application;

[0037] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] Below, embodiments of the present application will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the present application. In the following detailed description, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present application. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present application.

[0039] The terms used herein are only for describing specific embodiments and are not intended to limit the present application. The terms "include", "comprising", etc. used herein indicate the existence of the features, steps, operations and / or components, but do not exclude the existence or addition of one or more other features, steps, operations or components.

[0040] All terms (including technical and scientific terms) used herein have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification, and should not be interpreted in an idealized or overly rigid manner.

[0041] Some block diagrams and / or flow charts are shown in the accompanying drawings. It should be understood that some blocks or combinations thereof in the block diagrams and / or flow charts may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that these instructions, when executed by the processor, may create a device for implementing the functions / operations described in these block diagrams and / or flow charts.

[0042] In some embodiments, see Figure 1 , Figure 1 A schematic diagram of a process flow of a line insulator ultraviolet image data stream digital twin method provided in an embodiment of the present application; A line insulator ultraviolet image data stream digital twin method provided in an embodiment of the present application includes:

[0043] S110, obtaining observation data of sample insulators under different environments.

[0044] Insulators are affected to different degrees under different atmospheric environments. In lightly polluted areas with dryness, less fog or more rainfall, less dirt may accumulate on the surface of insulators; while in areas with humid fog but less rainfall, more dirt may accumulate. Therefore, the observation data may include the salt density and ash density values ​​on the surface of the insulator, as well as the changing trends of these values ​​over time and weather conditions. In addition, color coding may also be included. These observation data can reflect the surface texture characteristics of the insulator, which is convenient for the model to identify anomalies through surface texture. Here, different environments can include different observation times, such as summer, winter, morning, noon, and different lighting conditions, such as sunny, cloudy and rainy days. It is understandable that there are differences in the observation data corresponding to the sample insulators under different observation times and different lighting conditions, and no enumeration is performed for this. The sample insulators can be a large number of insulators that are easy to observe, and some of these insulators can be located in the same environment and some in different environments.

[0045] S120, performing digital twin modeling on the sample insulators in each type of environment based on the observed data to form at least two first digital twins.

[0046] Taking the lighting conditions as an example, the observation data of sample insulators on sunny days, cloudy days and rainy days are obtained. Based on these observation data, digital twin modeling is performed respectively, and the first digital twins corresponding to the sunny conditions, cloudy conditions and rainy conditions can be formed. It should be noted that the lighting conditions can also be distinguished by the specific light intensity. Here, the number of sample insulators in each environment should be appropriate to avoid accidental errors caused by too few numbers affecting the modeling accuracy, and to avoid too many numbers affecting the modeling efficiency. The specific number can be determined according to the actual accuracy requirements or experience, and there is no restriction on this.

[0047] S130, preprocessing the first digital twin, removing outliers in the first digital twin, and obtaining a second digital twin.

[0048] A digital twin, also known as a digital mapping or digital mirror, is a digital "clone" of a device or system created in a virtual space by making full use of data such as physical models, sensor updates, and operation history. Its biggest feature is the dynamic simulation of physical objects, which can reflect the entire life cycle of physical equipment in real time. In this embodiment, the first digital twin can be equivalent to a digital "clone" of the sample insulator. Outliers represent observations that are significantly different from other data points in the data set. By eliminating outliers in the first digital twin, accidental errors can be reduced and the fitting accuracy of the digital twin can be improved.

[0049] S140, using the second digital twin to train the initial neural network model to obtain a target neural network model.

[0050] Here, the second digital twin can be standardized and unified to the same coordinate system. After that, the standardized second digital twin can be divided into a training set, a validation set, and a test set, and the initial neural network model can be trained to obtain a target neural network model.

[0051] S150, inputting the insulator twin to be tested into the target neural network model to generate an evaluation result.

[0052] In this embodiment, the twin of the insulator to be tested is input into the target neural network model to obtain the prediction result of the insulator to be tested, and the prediction result is compared with the actual situation of the insulator to be tested, and an evaluation result is generated according to the difference between the two. If the difference between the two meets the preset conditions, it can be determined that the twin of the insulator to be tested is accurate and reliable. Otherwise, it can be determined that the twin of the insulator to be tested is inaccurate and unreliable, and a new twin needs to be created for the insulator to be tested.

[0053] In some embodiments, S130, preprocessing the first digital twin, removing outliers in the first digital twin, and obtaining the second digital twin includes:

[0054] Obtain the standard deviation between each data point in the first digital twin and the corresponding mean;

[0055] Determine data points whose standard deviation is greater than a preset threshold as outliers in the first digital twin;

[0056] The outliers are removed to obtain the second digital twin.

[0057] It is understandable that the presence of outliers may cause deviations in the parameter estimation of the first digital twin. Since the error of all data points needs to be minimized, and the presence of outliers may cause the model to overfit these outliers, resulting in inaccurate estimation of the overall data trend, it is necessary to remove the outliers. Here, outliers can be removed by filtering. That is, abnormal data points are screened out according to some statistical indicators or empirical rules. Statistical indicators may include standard deviations. It should be noted that statistical indicators may also include mean, median, percentile, etc., which are not limited. In addition, in addition to using filtering to remove outliers, other methods can also be used to achieve this, such as clustering, by grouping all data points into different clusters to identify outliers. Common clustering methods include K-means clustering and DBSCAN clustering. Another example is the anomaly detection algorithm, which uses algorithms specifically used to identify outliers, such as LOF algorithm, Isolation Forest algorithm and One-class SVM algorithm to identify outliers.

[0058] In some embodiments, S130 preprocesses the first digital twin, removes outliers in the first digital twin, and obtains the second digital twin, including:

[0059] The first digital twin is preprocessed using the ROR filter to remove outliers in the first digital twin and obtain the second digital twin.

[0060] The ROR (Robust Outlier Removal) filter can efficiently identify and remove noise and outliers to provide more accurate data. The principle of the ROR filter is based on statistical methods, calculating the distance between each point in the point cloud and its nearest neighbor to determine the degree of abnormality of the point. These distance values ​​can form a distribution, and then a threshold is set based on this distribution. Outliers are points whose distance values ​​exceed the threshold. Here, the specific value of the threshold can be set based on experience or accuracy requirements, and there is no limit on this.

[0061] In some embodiments, at S150, before inputting the insulator twin to be tested into the target neural network model and generating the evaluation result, the method further includes:

[0062] The insulator twins to be tested are preprocessed to remove outliers in the insulator twins to be tested.

[0063] Similar to the above embodiment, before the insulator twin to be tested is input into the target neural network model and the evaluation result is generated, it is also necessary to remove outliers in the insulator twin to be tested to ensure the accuracy of the insulator twin to be tested and avoid the influence of abnormal data on the evaluation result.

[0064] In some embodiments, the initial neural network model is a Grid-GCN model.

[0065] The Grid-GCN model includes an input-output module, a multi-scale grouping module, a local feature learning module, and a multi-scale feature fusion module. Its workflow can be summarized as follows: Input: The model accepts the original 3D point cloud as input. The point cloud is usually composed of a large number of discrete 3D coordinate points, and each point can be accompanied by additional attribute information (such as color, reflection intensity, etc.). Voxelization: Divide the 3D space into a regular grid structure (voxel grid), and each voxel represents a volume unit of fixed size. Then map each point in the point cloud to the corresponding voxel unit. Points in the same voxel form a "point group". Coverage-aware grid query: Identify empty voxels that do not contain any points; for points close to the edge of the voxel grid, especially those outside the voxels near the edge, the boundaries of the voxel grid need to be expanded to ensure that these points are also effectively covered. According to the actual distribution of the point cloud, the voxel size or grid layout is dynamically adjusted, especially in areas where the local density of the point cloud varies greatly, to ensure that each point is covered by at least one voxel, while minimizing the occurrence of invalid or sparse voxels. Graph convolution operation, each node calculates a new feature representation based on the features of its neighboring nodes, which can be achieved through weighted summation, splicing, attention mechanism, etc. The updated features of the node are further processed by a learnable weight matrix (similar to the convolution kernel in CNN) and a nonlinear activation function (such as ReLU). By superimposing multiple layers of GCA modules, the model can capture the multi-level and more complex dependencies of node features. Feature fusion and pooling, after the last layer of GCA, a global pooling operation (such as maximum pooling, average pooling, or learned pooling) may be used to summarize the features of all voxel nodes to obtain a compact representation of the entire point cloud; if the model contains multiple branches or multiple scales of features, cross-scale or multi-path feature fusion may be performed at this time to integrate feature information at different levels.

[0066] In some embodiments, S140, using the second digital twin to train the initial neural network model to obtain a target neural network model includes:

[0067] The initial neural network model is trained using the second digital twin and the corresponding real insulator model to obtain the target neural network model.

[0068] In some embodiments, a cross entropy loss function is used as a training loss function of the initial neural network model, and the training loss function satisfies:

[0069] ;

[0070] in, is the value of the i-th data point in the real insulator model, is the probability of the i-th category predicted by the target neural network model.

[0071] In this embodiment, the neural network model can be iteratively trained using the training set data. On each batch, the forward propagation data passes through the network, and the output and loss are calculated. The gradient is calculated using the backpropagation algorithm, and the weights and bias terms of the model are updated. After each training round, the performance of the model is evaluated using the validation set, and the performance indicators are recorded. When the prediction of the target neural network model is completely accurate, the cross entropy loss reaches a minimum value of 0.

[0072] In one example, the second digital twin sample set can be randomly divided into k parts, k-1 parts as training sets, and 1 part as validation set. The training set and validation set are rotated k times in turn, and the neural network model with the smallest validation error is selected as the target neural network model.

[0073] The embodiment of the present application obtains observation data of sample insulators in different environments, performs digital twin modeling on the sample insulators in each type of environment, and forms at least two first digital twins. After removing outliers in the first digital twins, the initial neural network model is trained using the digital twins to obtain a target neural network model. The target neural network model is used to judge the situation of the insulator twin to be tested, which fully considers the situation of the insulator in complex scenarios and facilitates accurate acquisition of the operating status of the line insulator.

[0074] In some embodiments, see Figure 2 , Figure 2 A schematic diagram of the structure of a line insulator ultraviolet image data stream digital twin device provided in an embodiment of the present application. A line insulator ultraviolet image data stream digital twin device 200 provided in an embodiment of the present application includes: an acquisition module 210, a creation module 220, a preprocessing module 230, a training module 240 and a generation module 250; wherein,

[0075] An acquisition module 210 is configured to acquire observation data of sample insulators under different environments;

[0076] A creation module 220 is configured to perform digital twin modeling on the sample insulators in each type of environment based on the observation data to form at least two first digital twins;

[0077] A preprocessing module 230 is configured to preprocess the first digital twin, remove outliers in the first digital twin, and obtain a second digital twin;

[0078] A training module 240 is configured to train the initial neural network model using the second digital twin to obtain a target neural network model;

[0079] The generation module 250 is configured to input the insulator twin to be tested into the target neural network model to generate an evaluation result.

[0080] In some embodiments, the pre-processing module 230 is specifically configured as follows:

[0081] Obtain the standard deviation between each data point in the first digital twin and the corresponding mean;

[0082] Determine data points whose standard deviation is greater than a preset threshold as outliers in the first digital twin;

[0083] The outliers are removed to obtain the second digital twin.

[0084] In some embodiments, the pre-processing module 230 is specifically configured as follows:

[0085] The first digital twin is preprocessed using the ROR filter to remove outliers in the first digital twin and obtain the second digital twin.

[0086] In some embodiments, the pre-processing module 230 is further configured to:

[0087] The insulator twins to be tested are preprocessed to remove outliers in the insulator twins to be tested.

[0088] In some embodiments, the initial neural network model is a Grid-GCN model.

[0089] In some embodiments, the training module 240 is specifically configured as follows:

[0090] The initial neural network model is trained using the second digital twin and the corresponding real insulator model to obtain the target neural network model.

[0091] In some embodiments, a cross entropy loss function is used as a training loss function of the initial neural network model, and the training loss function satisfies:

[0092] ;

[0093] in, is the value of the i-th data point in the real insulator model, is the probability of the i-th category predicted by the target neural network model.

[0094] The line insulator ultraviolet image data stream digital twin device provided in the embodiment of the present application can implement each process in the embodiment corresponding to the above-mentioned line insulator ultraviolet image data stream digital twin method. To avoid repetition, they will not be described here.

[0095] It should be noted that the line insulator ultraviolet image data stream digital twin device provided in the embodiment of the present application and the line insulator ultraviolet image data stream digital twin method provided in the embodiment of the present application are based on the same application concept. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned line insulator ultraviolet image data stream digital twin method, and the repeated parts will not be repeated.

[0096] In some embodiments, see Figure 3 , Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. An electronic device 300 provided in an embodiment of the present application includes a processor 310 and a memory 320; the memory 320 stores a computer program, wherein the computer program implements the above-mentioned line insulator ultraviolet image data stream digital twin method when executed by the processor.

[0097] Specifically, the processor 310 may include, for example, a general-purpose microprocessor, an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 310 may also include an onboard memory for cache purposes. The processor 310 may be a single processing unit or multiple processing units for executing different actions of the method flow according to the embodiment of the present application.

[0098] The memory 320 may be any medium capable of containing, storing, conveying, propagating or transmitting instructions. For example, the memory 320 may include, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared or semiconductor system, device, component or propagation medium. Specific examples of the memory 320 include: a magnetic storage device, such as a magnetic tape or a hard disk (HDD); an optical storage device, such as a compact disk (CD-ROM); a random access memory (RAM) or a flash memory; and / or a wired / wireless communication link.

[0099] The present application also provides a computer-readable medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned line insulator ultraviolet image data stream digital twin method. The computer-readable medium may be included in the device / apparatus / system described in the above-mentioned embodiment; or it may exist independently without being assembled into the device / apparatus / system. The above-mentioned computer-readable medium carries one or more programs, and when the above-mentioned one or more programs are executed, the method according to the embodiment of the present application is implemented.

[0100] According to an embodiment of the present application, a computer-readable medium may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that may be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to an electromagnetic signal, an optical signal, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wired, optical cable, radio frequency signal, etc., or any suitable combination of the above.

[0101] Those skilled in the art will appreciate that the features described in the various embodiments and / or claims of the present application may be combined in a variety of ways and / or combinations, even if such combinations or combinations are not explicitly described in the present application. In particular, without departing from the spirit and teachings of the present application, the features described in the various embodiments and / or claims of the present application may be combined in a variety of ways and / or combinations. All of these combinations and / or combinations fall within the scope of the present application. Therefore, the scope of the present application should not be limited to the above-described embodiments, but should be determined not only by the appended claims, but also by the equivalents of the appended claims.

Claims

1. A digital twin method for ultraviolet image data stream of line insulators, characterized in that: include: Obtain observation data of sample insulators under different environments; Based on the observation data, digital twin modeling is performed on the sample insulators in each type of environment to form at least two first digital twins; Preprocessing the first digital twin to remove outliers in the first digital twin to obtain a second digital twin; The initial neural network model is trained using the second digital twin and the corresponding real insulator model, and a cross entropy loss function is used as the training loss function of the initial neural network model to obtain a target neural network model; the training loss function satisfies: ;in, is the value of the i-th data point in the real insulator model, is the probability of the i-th category predicted by the target neural network model; The twin of the insulator to be tested is input into the target neural network model to generate an evaluation result; if the difference between the predicted result of the insulator to be tested output by the target neural network model and the actual situation meets the preset conditions, it is determined that the twin of the insulator to be tested is accurate and reliable; otherwise, it is determined that the twin of the insulator to be tested is inaccurate and unreliable.

2. The line insulator ultraviolet image data stream digital twin method according to claim 1 is characterized in that: The preprocessing of the first digital twin to remove outliers in the first digital twin to obtain a second digital twin includes: Obtaining a standard deviation between each data point in the first digital twin and the corresponding mean; Determine data points whose standard deviation is greater than a preset threshold as outliers in the first digital twin; The outliers are removed to obtain a second digital twin.

3. The line insulator ultraviolet image data stream digital twin method according to claim 1, characterized in that: The preprocessing of the first digital twin to remove outliers in the first digital twin to obtain a second digital twin includes: The first digital twin is preprocessed by using a ROR filter to remove outliers in the first digital twin to obtain a second digital twin.

4. The line insulator ultraviolet image data stream digital twin method according to claim 1, characterized in that: Before inputting the insulator twin to be tested into the target neural network model to generate an evaluation result, the method further includes: Preprocessing is performed on the insulator twins to be tested, and outliers in the insulator twins to be tested are eliminated.

5. The line insulator ultraviolet image data stream digital twin method according to claim 1, characterized in that: The initial neural network model is a Grid-GCN model.

6. A digital twin device for ultraviolet image data stream of line insulators, characterized in that: include: Get module, create module, preprocess module, train module and generate module; The acquisition module is configured to acquire observation data of sample insulators under different environments; The creation module is configured to perform digital twin modeling on the sample insulators in each type of environment based on the observation data to form at least two first digital twins; The preprocessing module is configured to preprocess the first digital twin, remove outliers in the first digital twin, and obtain a second digital twin; The training module is configured to train the initial neural network model using the second digital twin and the corresponding real insulator model, and adopt the cross entropy loss function as the training loss function of the initial neural network model to obtain the target neural network model; the training loss function satisfies: ;in, is the value of the i-th data point in the real insulator model, is the probability of the i-th category predicted by the target neural network model; The generation module is configured to input the insulator twin to be tested into the target neural network model to generate an evaluation result; if the difference between the predicted result of the insulator to be tested output by the target neural network model and the actual situation meets the preset conditions, it is determined that the insulator twin to be tested is accurate and reliable; otherwise, it is determined that the insulator twin to be tested is inaccurate and unreliable.

7. An electronic device comprising a processor and a memory; the memory stores a computer program, wherein: When the computer program is executed by the processor, the computer program implements the line insulator ultraviolet image data stream digital twin method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the steps of the method according to any one of claims 1 to 5 are implemented.

Images