Weather condition identification method and system for power transmission line, medium and equipment

Abstract

The invention provides a weather condition identification method and system for a power transmission line, a medium and equipment, and relates to the technical field of power equipment operation and maintenance and weather forecast, and the weather condition identification method for the power transmission line comprises the steps: respectively obtaining image data and numerical data corresponding to an environment where the power transmission line is located; inputting the image data into a first neural network model to generate a feature map corresponding to the image data; inputting the numerical data into a second neural network model to generate a numerical sequence corresponding to the numerical data; based on a bidirectional cross attention mechanism, performing fusion processing on the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data to generate a fusion feature vector; based on the fusion feature vector, the weather condition corresponding to the environment where the power transmission line is located is identified, and the accuracy of weather state identification is improved.

Description

Technical Field

[0001] This application generally relates to the fields of power system operation and maintenance and weather forecasting technology. More specifically, this application relates to a method, system, medium, and device for weather condition identification of transmission lines. Background Technology

[0002] As a critical infrastructure in the power system, the safety and stability of power transmission lines directly affect the normal operation of the national economy and the order of social production and life. However, due to geographical environment and climate conditions, many transmission lines are exposed to remote and desolate outdoor environments for extended periods, where weather conditions are variable and they are susceptible to severe weather such as snowfall, freezing rain, and heavy fog. This can lead to safety hazards such as conductor icing, equipment failure, and even large-scale power outages, making it difficult for maintenance personnel to access these areas. Therefore, timely and accurate identification of the meteorological conditions in the areas where transmission lines are located is of great significance for ensuring the safe and stable operation of the power grid.

[0003] Image data possesses spatially structured features (such as the blurred texture of fog and the particle outline of snow), while numerical data possesses dynamically correlated features (such as the risk of line icing due to temperatures <-2℃ and humidity >85%). The semantic dimensions of these two types of features differ significantly. Existing technologies employ "stitching" or "one-way attention" methods, failing to achieve semantic alignment between image and numerical data. The feature vector obtained after their fusion contains redundancy or loses key information, leading to inaccurate recognition results (e.g., the blurred features of a foggy image cannot be correlated with the numerical features of high humidity, resulting in a misclassification as a sunny day).

[0004] Because the meteorological conditions in the environment where power transmission lines are located are sudden (such as freezing rain falling suddenly on a sunny day) and fluctuating (such as a sudden drop in visibility due to the spread of dense fog), the fusion process of image data and numerical data is subject to a fixed fusion ratio in the existing technology. This makes it impossible to adapt to changes in the effectiveness of features in real time (such as the image features becoming invalid in dense fog, but still being fused at a 50% ratio, resulting in the recognition accuracy dropping from 90% to below 70%). The lack of dynamic adjustment capability leads to low recognition accuracy. Summary of the Invention

[0005] In order to at least address one or more of the technical problems mentioned above, this application proposes a method, system, medium, and device for weather condition identification of transmission lines in several aspects.

[0006] In the first aspect, the weather condition identification method for transmission lines provided in this application includes the following steps: Image data and numerical data corresponding to the environment where the transmission line is located are acquired respectively. The image data includes at least multiple images of the line equipment area and multiple images of the background area. The numerical data includes at least temperature, humidity, wind speed and visibility. The image data is input into a first neural network model to generate a feature map corresponding to the image data; The numerical data is input into the second neural network model to generate a numerical sequence corresponding to the image data; Based on the bidirectional cross-attention mechanism, the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data are fused to generate a fused feature vector. Based on the fused feature vector, the weather conditions corresponding to the environment in which the transmission line is located are identified.

[0007] In some examples, the method further includes, before inputting the image data into the first neural network model: The image data is input into a pre-trained line equipment detection model to generate a mask matrix corresponding to the image data, wherein the value corresponding to the line equipment region image is 1 and the value corresponding to the background region image is 0 in the mask matrix. The mask matrix corresponding to the image data is embedded into the feature map corresponding to the image data.

[0008] In some examples, the method further includes, before inputting the numerical data into the second neural network model: The numerical data is classified according to the weather type to obtain multiple sets of data, wherein the multiple sets of data include numerical data and corresponding weather types; Based on the magnitude of the numerical data in the multiple sets of data, assign corresponding weights to the weather types.

[0009] In some examples, the interactive processing of the feature map and the numerical sequence to generate the second cross-attention weights corresponding to the feature map and the numerical sequence includes: According to the formula Bidirectional cross-attention calculations are performed on the feature map and the numerical sequence respectively to generate a first cross-attention weight for the feature map and a second cross-attention weight for the numerical sequence, wherein... For the cross-attention weights, Softmax() is the normalized exponential function, Q is the query vector, K is the key vector, V is the value vector, and Sim is the mask matrix. Sim is used to calibrate the attention weights.

[0010] In some examples, after inputting the image data into a first neural network model to generate a feature map corresponding to the image data, the method further includes: The feature map is compressed so that the number of channels in the feature map is consistent with the embedding dimension of the second neural network model.

[0011] In some examples, after generating the numerical sequence corresponding to the numerical data, the method further includes: The first feature vector corresponding to the feature map is dimensionally reduced so that the first feature vector is dimensionally aligned with the second feature vector corresponding to the numerical sequence.

[0012] In some examples, the first neural network model is a ConvNeXt model.

[0013] In some examples, the second neural network model is a Transformer model.

[0014] In a second aspect, the weather condition identification system for transmission lines provided in this application includes: The acquisition module is configured to acquire image data and numerical data corresponding to the environment where the transmission line is located, wherein the image data includes at least multiple images of the line equipment area and multiple images of the background area, and the numerical data includes at least temperature, humidity, wind speed and visibility. The generation module is configured to input the image data into a first neural network model and generate a feature map corresponding to the image data. The generation module is further configured to input the numerical data into a second neural network model to generate a numerical sequence corresponding to the numerical data. The fusion module is configured to perform fusion processing on the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data based on a bidirectional cross-attention mechanism to generate a fused feature vector. The identification module is configured to identify the weather conditions corresponding to the environment in which the transmission line is located based on the fused feature vector.

[0015] In a third aspect, this application provides a computer-readable storage medium containing program instructions that, when executed by a processor, cause the method described in the first aspect to be implemented.

[0016] In a fourth aspect, this application provides an electronic device, comprising: Processor; and A memory that stores computer instructions that, when executed by the processor, cause the electronic device to perform the method described in the first aspect above.

[0017] By adopting the above technical solution, this application has the following beneficial effects: By generating a fusion feature vector through a collaborative bidirectional attention fusion mechanism involving visual and numerical agents, weather conditions can be identified based on this fusion feature vector, thereby improving the accuracy of weather condition identification. Attached Figure Description

[0018] The above and other objects, features, and advantages of exemplary embodiments of this application will become readily understood by reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of this application are illustrated by way of example and not limitation, and the same or corresponding reference numerals denote the same or corresponding parts, wherein: Figure 1 An exemplary flowchart of a weather condition identification method for power transmission lines provided in an embodiment of this application is shown. Figure 2 A schematic diagram of an exemplary framework for a weather condition identification system for power transmission lines, provided in an embodiment of this application, is shown.

[0019] Figure 3 An exemplary structural block diagram of an electronic device according to some embodiments of this application is shown. Detailed Implementation

[0020] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0021] It should be understood that the terms "comprising" and "including" used in the specification and claims of this application indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0022] It should also be understood that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the application. As used in this specification and claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this specification and claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes such combinations.

[0023] As used in this specification and claims, the term "if" may be interpreted, depending on the context, as "when," "once," "in response to determination," or "in response to detection." Similarly, the phrase "if determined" or "if [described condition or event] is detected" may be interpreted, depending on the context, as "once determined," "in response to determination," "once [described condition or event] is detected," or "in response to detection of [described condition or event]."

[0024] The specific embodiments of this application will now be described in detail with reference to the accompanying drawings.

[0025] Example 1 like Figure 1 As shown, the weather condition identification method for transmission lines provided in this application includes the following steps: S101, acquire image data and numerical data corresponding to the environment where the transmission line is located, wherein the image data includes at least an image of the line equipment area and an image of the background area, and the numerical data includes at least temperature, humidity, wind speed and visibility.

[0026] Specifically, high-definition camera modules fixedly installed on the towers where the transmission lines are located can collect image data of the environment corresponding to the environment where the transmission lines are located, and various sensor modules (temperature sensors, humidity sensors, etc.) fixedly installed on the towers where the transmission lines are located can collect numerical data of the environment corresponding to the environment where the transmission lines are located.

[0027] S102, input the image data into the first neural network model to generate the feature map corresponding to the image data.

[0028] In some examples, the method further includes the following steps prior to this one: The image data is input into a pre-trained line equipment detection model to generate a mask matrix corresponding to the image data. In this mask matrix, the value corresponding to the line equipment area image is 1, and the value corresponding to the background area image is 0. Embed the mask matrix corresponding to the image data into the feature map corresponding to the image data.

[0029] In this process, after the image data is input into the first neural network model, the model uses a mask matrix to weight and enhance the equipment area image (weight coefficient 1.5) while suppressing the background image (weight coefficient 0.3). This allows the first neural network model to prioritize the extraction of features related to weather conditions in the equipment area image (such as ice thickness on conductors and fog concentration around towers), providing a reliable foundation for accurate subsequent weather condition identification. This first neural network model is a ConvNeXt model, and its output feature map... Where R is a real number, B1 is the number of images processed at once, C is the number of channels in the feature map, H is the height of the image data, and W is the width of the image data. The ConvNeXt model, as a modern convolutional neural network, achieves a good balance between performance and efficiency, making it suitable for real-time weather recognition tasks. This network consists of multiple convolutional stages, with the spatial resolution of the image gradually decreasing and the semantic representation gradually increasing in each stage. Each stage introduces depthwise separable convolutions with large kernels to expand the receptive field, employs an inverse residual structure to enhance nonlinear expressive power, and stabilizes the training process through layer normalization. This hierarchical design enables the visual agent to simultaneously extract local texture and global contextual information, thereby supporting an accurate understanding of fine-grained weather conditions. The final feature map has 768 channels.

[0030] In some examples, after step S102, the method further includes: The feature map is compressed so that the number of channels in the feature map is consistent with the embedding dimension of the second neural network model.

[0031] Specifically, a 1×1 convolution kernel is used to compress the number of channels in the feature map, reducing the number of channels in the feature map from C to D, so that it can be interacted and fused with numerical data in the same embedding space.

[0032] S103, input the numerical data into the second neural network model to generate the numerical sequence corresponding to the numerical data.

[0033] Specifically, the second neural network model is a Transformer model, and its output is a numerical sequence. Where R is a real number, B² is the number of values ​​processed at once, S is the number of value categories, and D is the embedding dimension of the second neural network model. The Transformer model is used to model the complex relationships between different environmental variables. Each input feature is first projected into a high-dimensional embedding space through a shared linear transformation, and positional encoding is added to distinguish the location information of each feature. Subsequently, these feature embeddings are processed by multiple Transformer encoder layers, achieving global interaction between variables through a self-attention mechanism, enabling the model to learn feature combinations that are discriminative about weather conditions.

[0034] In some examples, the method further includes, before inputting the numerical data into the second neural network model: The numerical data is classified according to the weather type to obtain multiple sets of data, wherein the multiple sets of data include numerical data and corresponding weather types; Based on the magnitude of the numerical data in the multiple sets of data, assign corresponding weights to the weather types.

[0035] Specifically, variables such as temperature, humidity, wind speed, and visibility are categorized according to weather conditions (e.g., key parameters for icing: temperature, humidity, and wind speed; key factors for fog: humidity and visibility), and weights are assigned to each weather condition (e.g., the weight for icing is increased to 1.8 when the temperature is <-2℃). This allows the Transformer model to prioritize combinations of variables strongly correlated with the target weather. For combinations of irrelevant variables (e.g., wind speed and fog are weakly correlated), a corresponding mask matrix is ​​set to prevent excessive interaction (attention weight is set to 0), allowing only strongly correlated variables (e.g., temperature and humidity) to interact globally. By combining meteorological knowledge with the self-attention mechanism of the Transformer model, indiscriminate interaction between numerical values ​​is avoided, improving feature recognition efficiency.

[0036] In some examples, after step S103, the method further includes: The first feature vector corresponding to the feature map is dimensionally reduced so that the first feature vector is dimensionally aligned with the second feature vector corresponding to the numerical sequence.

[0037] S104, based on the bidirectional cross-attention mechanism, the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data are fused to generate a fused feature vector.

[0038] In some examples, this step specifically includes: According to the formula (1) Bidirectional cross-attention calculations are performed on the feature map and the numerical sequence respectively to generate the first cross-attention weights for the feature map and the second cross-attention weights for the numerical sequence. In equation (1), For the cross-attention weights, Softmax() is the normalized exponential function, Q is the query vector, K is the key vector, V is the value vector, and Sim is the mask matrix. Sim is used to calibrate the attention weights and avoid invalid interactions (such as the low similarity between the feature map corresponding to sunny days and the value of high humidity, resulting in suppressed weights). This makes the feature map and the numerical sequence semantically aligned, improves the fusion effect, and provides more accurate feature support for the final weather state recognition.

[0039] Specifically, after obtaining the first and second cross-attention weights, pooling operations are performed on the first and second cross-attention weights along the sequence dimension, as follows: (2) (3) In equations (2) and (3), The first cross-attention weights, The second cross-attention weight; The result obtained from the first cross-attention weight pooling operation The result obtained from the second cross-attention weight pooling operation After concatenation, the data is input into a gating network to generate fusion gating weights g. These weights adaptively adjust the contribution ratio of various feature data in the fusion process, achieving dynamic feature fusion. The calculation process is as follows: (4) In equation (4), σ represents the Sigmoid activation function, W g W r The weights are learnable parameters used to implement linear transformations in the gated network, and ReLU() is the activation function.

[0040] Finally, the feature vector F is fused. f It is calculated in the following way: (5) In equation (5), the symbol This indicates that the elements are multiplied one by one.

[0041] This fusion strategy enables the full utilization of complementary information from the first neural network model (visual agent) and the second neural network model (numerical agent), thereby achieving more accurate and robust prediction results.

[0042] S105. Based on the fused feature vector, identify the weather conditions corresponding to the environment in which the transmission line is located.

[0043] Specifically, the fused feature vector is further processed by the prediction head of the Transformer model, and finally outputs the probability distribution of each category of weather conditions, thereby realizing the identification of weather conditions.

[0044] Among them, a fusion feature vector is generated by a two-way attention fusion mechanism of visual intelligent agent and numerical intelligent agent, and the weather state is identified based on the fusion feature vector to achieve accurate weather state identification.

[0045] To verify the validity of this application, a series of comparative experiments were conducted. As shown in Table 1, in these experiments, the performance of the multi-agent model used in this application was compared with that of the single-agent baseline model using only image modality or only numerical modality.

[0046] Table 1 plan accuracy Accuracy Recall rate F1 score Visual intelligent agents 93.19 92.32 92.80 92.53 Numerical intelligent agents 90.23 93.24 86.49 88.91 This application 95.12 94.24 95.25 94.73 Experimental results show that the proposed method comprehensively outperforms the two single-agent methods, achieving the optimal performance with an accuracy of 95.12% and an F1 score of 94.73%, demonstrating superior overall performance. This result confirms that the method proposed in this application, which integrates visual and numerical modalities using a multi-agent framework, can effectively improve the accuracy of weather condition recognition.

[0047] To further verify the validity of this application, a comparative experiment was conducted with a representative single agent, and the results are shown in Table 2.

[0048] Table 2 Experimental data show that our application outperforms all compared single-agent systems in four metrics: accuracy, precision, recall, and F1 score. This result highlights the advantages of multimodal information fusion and confirms that our application has superior feature representation capabilities and stronger generalization performance in weather condition recognition tasks.

[0049] Example 2 like Figure 2 As shown, the weather condition identification system for transmission lines provided in this application includes: The acquisition module is configured to acquire image data and numerical data corresponding to the environment where the transmission line is located, wherein the image data includes at least multiple images of the line equipment area and multiple images of the background area, and the numerical data includes at least temperature, humidity, wind speed and visibility. The generation module is configured to input the image data into a first neural network model and generate a feature map corresponding to the image data. The generation module is further configured to input the numerical data into a second neural network model to generate a numerical sequence corresponding to the numerical data. The fusion module is configured to perform fusion processing on the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data based on a bidirectional cross-attention mechanism to generate a fused feature vector. The identification module is configured to identify the weather conditions corresponding to the environment in which the transmission line is located based on the fused feature vector.

[0050] On the other hand, embodiments of this application also provide an electronic device, see [link to relevant documentation]. Figure 3 , Figure 3 This is an exemplary structural block diagram of an electronic device according to an embodiment of this application, such as... Figure 3 As shown, the electronic device includes a processor and a memory, the memory storing computer instructions, and the processor executing the computer instructions to perform the method provided in this application.

[0051] Specifically, processor 601 may include a central processing unit (CPU) or a graphics processing unit (GPU), or an application-specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application. Memory 602 may include memory for data or instructions. For example, memory 602 may be at least one of the following: a hard disk drive (HDD), read-only memory (ROM), random access memory (RAM), floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, universal serial bus (USB) drive, or other physical / tangible memory storage device. Alternatively, memory 602 may include removable or non-removable (or fixed) media. Furthermore, memory 602 may be internal or external to the integrated gateway disaster recovery device. Memory 602 may be non-volatile solid-state memory. In other words, typically memory 602 includes a tangible (non-transitory) computer-readable storage medium (such as a memory device) encoded with executable instructions, wherein the stored executable instructions, when executed by processor 601 (e.g., by one or more processors), can implement the methods in the embodiments of this application.

[0052] In one example Figure 3 The illustrated electronic device may also include a communication interface 603 and a bus 610. The processor 601, memory 602, and communication interface 603 are connected via bus 610 and communicate with each other. Communication interface 603 is primarily used to enable communication between modules, devices, units, and / or equipment within the electronic device. Bus 610, including hardware, software, or both, couples components of the online data flow metering device together. For example, the bus may include at least one of the following: Accelerated Graphics Port (AGP) or other graphics bus, Enhanced Industry Standard Architecture (EISA) bus, Front Side Bus (FSB), HyperTransport (HT) interconnect, Industry Standard Architecture (ISA) bus, Infinite Bandwidth Interconnect, Low Pin Count (LPC) bus, memory bus, Microchannel Architecture (MCA) bus, Peripheral Component Interconnect (PCI) bus, PCI-Express (PCI-X) bus, Serial Advanced Technology Attachment (SATA) bus, Video Electronics Standards Association Local (VLB) bus, or other suitable buses. Bus 610 may include one or more buses. Although specific buses are described or illustrated in the embodiments of this application, any suitable bus or interconnection method may be considered in the embodiments of this application.

[0053] In another aspect, embodiments of this application also provide a computer-readable storage medium storing computer program instructions that, when executed by a processor, implement the aforementioned method. The computer-readable storage medium may be, for example, a classic computer-readable storage medium, such as a read-only memory (ROM), random access memory (RAM), disk storage media, optical storage media, flash memory, or other electrical, optical, or other physical / tangible memory storage devices.

[0054] In another aspect, embodiments of this application also provide a computer program product, which includes computer program instructions that, when executed by a processor, implement the method provided in embodiments of this application. This computer program product may be, for example, a software installation package, a plug-in compatible with a related software system, etc.

[0055] The flowcharts and / or block diagrams of the methods and systems of embodiments of this application have been described above by way of example, and related aspects have been described. It should be understood that each block or combination thereof in the flowcharts and / or block diagrams can be implemented by computer program instructions, by dedicated hardware performing a specified function or action, or by a combination of dedicated hardware and computer instructions. When implemented in hardware, it can be, for example, an electronic circuit, an application-specific integrated circuit (ASIC), appropriate firmware, a plug-in, a function card, etc.; when implemented in software, it is a program or code segment used to perform the required task. The program or code segment can be stored in memory or transmitted over a transmission medium or communication link via data signals carried in a carrier wave. The code segment can be downloaded via a computer network such as the Internet, an intranet, etc.

[0056] It should be noted that in this application, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0057] While this application has shown and described numerous embodiments, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will arise for those skilled in the art without departing from the spirit and intent of this application. It should be understood that various alternatives to the embodiments of this application described herein may be employed in the practice of this application. The appended claims are intended to define the scope of protection of this application and therefore cover equivalents or alternatives within the scope of these claims.

Claims

1. A weather condition identification method for a power transmission line, comprising: obtaining image data and numerical data corresponding to an environment in which the power transmission line is located, wherein the image data at least includes a plurality of line equipment region images and a plurality of background region images, and the numerical data at least includes temperature, humidity, wind speed and visibility; inputting the image data into a first neural network model to generate a feature map corresponding to the image data; inputting the numerical data into a second neural network model to generate a numerical sequence corresponding to the numerical data; based on a bidirectional cross-attention mechanism, performing fusion processing on the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data to generate a fusion feature vector; based on the fusion feature vector, identifying a weather condition corresponding to the environment in which the power transmission line is located.

2. The weather condition recognition method according to claim 1, characterized in that, Before inputting the image data into the first neural network model, the method further comprises: inputting the image data into a pre-trained line equipment detection model to generate a mask matrix corresponding to the image data, wherein in the mask matrix, the value corresponding to the line equipment region image is 1, and the value corresponding to the background region image is 0; embedding the mask matrix corresponding to the image data into the feature map corresponding to the image data.

3. The weather condition recognition method according to claim 1, characterized in that, Before inputting the numerical data into the second neural network model, the method further comprises: classifying the numerical data according to weather types to obtain a plurality of groups of data, wherein the plurality of groups of data include numerical data and corresponding weather types; setting corresponding weights for the corresponding weather types according to the size of the numerical data in the plurality of groups of data.

4. The weather condition recognition method according to claim 2, characterized in that, The interaction processing on the feature map and the numerical sequence to generate a second cross-attention weight corresponding to the feature map and the numerical sequence comprises: According to the formula , bidirectional cross-attention calculation is performed on the feature map and the numerical sequence respectively to generate first cross-attention weights of the feature map and second cross-attention weights corresponding to the numerical sequence, wherein, is the cross-attention weight, Softmax() is a normalization exponential function, Q is a query vector, K is a key vector, V is a value vector, Sim is a mask matrix, and Sim is used to calibrate the attention weight.

5. The weather condition recognition method according to claim 1, characterized in that, After inputting the image data into the first neural network model to generate the feature map corresponding to the image data, the method further comprises: compressing the feature map so that the number of channels of the feature map is consistent with the embedding dimension of the second neural network model.

6. The weather condition recognition method according to claim 5, characterized in that, After generating the numerical sequence corresponding to the numerical data, the method further comprises: performing dimension reduction on a first feature vector corresponding to the feature map so that the first feature vector is aligned in dimension with a second feature vector corresponding to the numerical sequence.

7. The weather condition recognition method according to claim 1, characterized in that, The first neural network model is a ConvNeXt model.

8. The weather condition recognition method according to claim 1, characterized in that, The second neural network model is a Transformer model.

9. A weather condition identification system for a power transmission line, comprising: an obtaining module configured to obtain image data and numerical data corresponding to an environment in which the power transmission line is located, wherein the image data at least includes a plurality of line equipment region images and a plurality of background region images, and the numerical data at least includes temperature, humidity, wind speed and visibility; a generating module configured to input the image data into a first neural network model to generate a feature map corresponding to the image data; the generating module is further configured to input the numerical data into a second neural network model to generate a numerical sequence corresponding to the numerical data; The fusion module is configured to perform fusion processing on the feature map corresponding to the image data and the numerical sequence corresponding to the numerical data based on a bidirectional cross-attention mechanism to generate a fusion feature vector. The identification module is configured to identify a weather condition corresponding to an environment where the power transmission line is located based on the fusion feature vector.

10. A computer-readable storage medium, characterized in that, Program instructions are included, and when the program instructions are executed by a processor, a method according to any one of claims 1-8 is implemented.

11. An electronic device, comprising: Comprise: A processor; And A memory storing computer instructions, when the computer instructions are run by the processor, the electronic device executes the method according to any one of claims 1-8.

Images