Power inspection method and system based on satellite remote sensing and large model

Abstract

The application relates to the technical field of power inspection, in particular to a power inspection method and system based on satellite remote sensing and a large model, which comprises the following steps: acquiring satellite remote sensing images of a target area; judging whether there is an abnormal feature in the target area based on the satellite remote sensing images; if there is, determining an inspection task and planning an inspection route based on the number and location of the abnormal feature; acquiring a business intelligent agent that can be used to execute the inspection task at present; and controlling the business intelligent agent to travel along the planned inspection route and complete the inspection task. The scheme provided by the application can improve the inspection efficiency of a remote target and reduce the inspection cost of the remote target.

Description

Technical Field

[0001] This application relates to the field of power line inspection technology, specifically to a power line inspection method, system, processor, and machine-readable storage medium based on satellite remote sensing and large models. Background Technology

[0002] In the power industry, the inspection of transmission channels is a crucial link in ensuring the safe and stable operation of the power grid. Traditional inspection methods mainly rely on manual labor, but this method has many limitations, such as high labor costs, low inspection efficiency, and significant safety risks. Especially in complex environments, such as mountainous areas and uninhabited areas, the difficulty and danger of manual inspection are further increased. Summary of the Invention

[0003] The purpose of this application is to provide a power line inspection method and system based on satellite remote sensing and large models, so as to reduce the cost of remote target inspection and improve inspection efficiency.

[0004] To achieve the above objectives, the first aspect of this application provides a power line inspection method based on satellite remote sensing and a large model. The method includes: S1, acquiring satellite remote sensing images of a target area; S2, determining whether there are abnormal features in the target area based on the satellite remote sensing images; S3, if there are, determining the inspection task and planning the inspection route based on the number and location of the abnormal features; S4, acquiring the business intelligence agent currently available to perform the inspection task; S5, controlling the business intelligence agent to move along the planned inspection route to complete the inspection task.

[0005] Based on the first aspect, in the embodiments of this application, if there are multiple inspection tasks, before controlling the business intelligence agent to proceed according to the planned inspection route, the method further includes: determining the priority of the inspection tasks; matching business intelligence agents for each inspection task in descending order of priority; traversing all inspection tasks until all inspection tasks are matched with the corresponding business intelligence agents.

[0006] Based on the first aspect, in the embodiments of this application, if there are multiple inspection tasks, before controlling the business intelligence agent to proceed according to the planned inspection route, the method further includes: matching the inspection tasks with the business intelligence agent in a many-to-one or one-to-one manner with the shortest inspection time as the objective.

[0007] Based on the first aspect, in the embodiments of this application, the method further includes: analyzing the equipment requirements of the inspection task for external plug-in devices; establishing business intelligent agents for different inspection tasks, and adding the necessary external plug-in devices to different business intelligent agents.

[0008] Based on the first aspect, in the embodiments of this application, the method further includes: acquiring user instructions; the user instructions include text instructions or voice instructions; parsing the user instructions to determine user inspection needs; and formulating inspection tasks based on user inspection needs.

[0009] Based on the first aspect, in the embodiments of this application, parsing user instructions and determining user inspection needs includes: writing inspection needs according to a pre-established prompt word template; obtaining prompt words from user instructions; and identifying user inspection needs based on prompt words and context.

[0010] Based on the first aspect, in the embodiments of this application, the business intelligent agent includes an unmanned camera.

[0011] Secondly, this application provides a power line inspection system based on satellite remote sensing and a large model. The system includes: a satellite module for acquiring satellite remote sensing images of a target area; a large model module for determining whether there are abnormal features within the target area based on the satellite remote sensing images; when abnormal features are present, it is also used to determine inspection tasks and plan inspection routes based on the number and location of the abnormal features; a status monitoring module for monitoring the working status of each business agent and acquiring the business agents currently available to perform inspection tasks; and business agents for performing inspection tasks by traveling along the planned inspection routes.

[0012] Thirdly, this application provides a processor configured to execute the aforementioned power line inspection method based on satellite remote sensing and large models.

[0013] Fourthly, this application provides a machine-readable storage medium storing instructions that, when executed by a processor, configure the processor to perform the aforementioned power line inspection method based on satellite remote sensing and a large model.

[0014] The solution provided in this application has at least the following beneficial effects: The solution provided in this application, on the one hand, rapidly acquires satellite remote sensing image data of the target area through satellite remote sensing, solving the problem of information source and reducing manpower consumption; on the other hand, based on the image and language understanding and task planning and allocation capabilities of a pre-trained large model, it can quickly determine inspection targets, generate inspection tasks, and call the corresponding business intelligence agent to execute tasks according to the (equipment) requirements of the inspection tasks, thereby improving the intelligence and automation level of inspection, increasing inspection efficiency, and reducing the maintenance cost of target equipment in the target area.

[0015] Other features and advantages of the embodiments of this application will be described in detail in the following detailed description section. Attached Figure Description

[0016] The accompanying drawings are provided to further illustrate the embodiments of this application and form part of the specification. They are used together with the following detailed description to explain the embodiments of this application, but do not constitute a limitation on the embodiments of this application. In the drawings: Figure 1 A flowchart illustrating a power line inspection method based on satellite remote sensing and a large model is shown. Figure 2 The schematic diagram illustrates a flowchart of an inspection task allocation method in one embodiment of this application; Figure 3 The diagram illustrates the structure of a power line inspection system based on satellite remote sensing and a large model.

[0017] Explanation of reference numerals in the attached figures 1-Satellite module; 2-Large model module; 3-Status monitoring module; 4-Business intelligent agent. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are only for illustration and explanation of the embodiments of this application and are not intended to limit the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0019] It should be noted that if the embodiments of this application involve directional indicators (such as up, down, left, right, front, back, etc.), the directional indicators are only used to explain the relative positional relationship and movement of each component in a certain specific posture (as shown in the figure). If the specific posture changes, the directional indicators will also change accordingly.

[0020] Furthermore, if the embodiments of this application involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, features defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the technical solutions of various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. If the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed in this application.

[0021] Example 1

[0022] like Figure 1As shown in the figure, this embodiment provides a power line inspection method based on satellite remote sensing and a large model. The method includes: S1. Acquire satellite remote sensing images of the target area; S2. Determine whether there are any abnormal features within the target area based on the satellite remote sensing image; S3. If they exist, the inspection task and the inspection route are determined based on the number and location of the abnormal features. S4. Obtain the business intelligence agents currently available to perform inspection tasks; S5. Control the business intelligence agent to move along the planned inspection route and complete the inspection task.

[0023] Specifically, in step S1, satellite remote sensing technology can be used to acquire satellite remote sensing images of the target area, such as high-frequency satellite remote sensing technology (e.g., Jilin-1, Planet constellations), which has high resolution (sub-meter level) and high revisit rate (multiple times per day). For example, if the target area includes power transmission channels and their surrounding areas, given the long and narrow nature of power transmission channels, satellite remote sensing can greatly improve inspection efficiency compared to manual inspection and detect potential external damage. A single satellite can cover a power transmission channel hundreds of kilometers long in a single transit, offering unparalleled efficiency advantages in detecting large-scale, wide-ranging potential external damage compared to manual or drone inspections. In addition, it can also detect macroscopic changes that are difficult to detect from the ground or low altitude, such as newly built roads, expanding construction sites, mining areas, and large-scale deforestation or vegetation growth.

[0024] Steps S2 and S3 can be completed using large-scale model technology (such as multimodal large models (LMM) or vertical domain-trained visual large models (VLM). This involves pre-training a machine learning model with large-scale parameters and computational power, enabling it to have dialogue, generation, and image recognition capabilities. The dialogue capability allows it to receive and understand user business needs through multi-turn dialogue; the generation capability can be used to generate text content such as inspection reports and analysis reports; and the image recognition capability can be used for remote sensing image recognition, particularly for identifying potential external force damage hazards (i.e., abnormal features) in remote sensing images of power transmission channels. For step S2, the large model can perform end-to-end intelligent analysis of the images acquired in step S1. This includes accurately identifying various potential external force damage hazards (abnormal features), such as large-scale construction... Construction machinery, illegally constructed buildings / ponds, excessively tall trees (tree obstructions), risks of wildfires / landslides / collapses, engineering blasting points, etc. It can also perform change detection, for example, by comparing with images from a previous time period (such as one week ago, one month ago), automatically detecting and marking areas where changes have occurred, greatly improving the efficiency of discovering new hazards. For example, maintenance personnel can interact with the large model using natural language, such as: "Find all possible cranes in the A12-A23 section of the passage in the last 3 days." After understanding the command, the large model automatically calls the corresponding analysis module and provides the results to the user in a visually appealing format with highlighted annotations. For step S3, the large model automatically generates a specific inspection task list based on the hazard list output in step S2. For example: "Task 1: Go to coordinates [XXX, [YYY] Confirm the crane type and its clearance distance from the conductor; Task 2: Head to coordinates [ZZZ, AAA] to assess the potential risks of the new plant to the power line. The large model can intelligently recommend task priorities and handling suggestions based on the type of hazard, historical data, and power grid regulations. Then, the large model takes the coordinates and priority of the task point as input and calls the built-in path planning algorithm interface (such as OR-Tools). The algorithm will generate a collaborative optimal path for the drone swarm executing the task, ensuring the shortest total time, no omissions, and compliance with the drone's endurance, airspace, and other constraints.

[0025] In a specific embodiment, the method of determining whether there are abnormal features in a target area based on the satellite remote sensing image in this application specifically includes: obtaining the resolution R of the satellite remote sensing image; calculating the signal-to-noise ratio (SNR) of the target area image; and determining that there are abnormal features if the SNR is greater than a set threshold_SNR.

[0026] In another embodiment, the total number of abnormal points N is counted; and a priority factor P = (D / max_D) × (N / max_N) is calculated based on the distance D between the abnormal point and the power supply station, where max_D and max_N are the maximum distance and the maximum number of abnormal points, respectively; and an inspection order is generated based on the P value.

[0027] In practice, the first step is to count the total number of anomalies, N. N reflects the frequency and scope of the faults; the larger the value, the more areas require priority handling. For example, in a large power transmission network, N may be as high as several hundred or even thousands.

[0028] Then, the priority factor P is calculated based on the distance D between the anomaly and the power station: P = (D / max_D) × (N / max_N). Here, D represents the distance from a single anomaly to the nearest power station, max_D is the farthest distance among all anomalies, N is the current number of anomalies, and max_N is the maximum number of anomalies in all samples. The P value ranges from 0 to 1; when P is close to 1, it indicates that the anomaly has a high priority and should be addressed as soon as possible. By combining distance and quantity in the calculation, both the urgency of the geographical location and the severity of the problem are considered. Finally, the inspection order is generated based on the P values. This step arranges the anomalies according to their P values ​​from high to low, forming an inspection priority list. For example, an area that is far away but has multiple anomalies will have a higher P value than nearby but fewer anomalies, thus being prioritized for inspection.

[0029] Choosing the optimal route reduces travel distance and improves inspection efficiency. Path planning algorithms, such as shortest path or traveling salesman problem solutions, ensure that inspectors can visit all high-priority anomaly points in the shortest possible time. Specifically, optimizing driving routes using geographic information data effectively reduces inspection time and costs. This method scientifically and rationally allocates limited inspection resources, avoids resource waste, and improves the efficiency and safety of power system operation and maintenance.

[0030] Furthermore, this application also determines the weighting coefficient α_i for the risk impact of different types of anomalies. The parameter α_i is used to quantify the risk level of each anomaly type, and its value is typically between 0 and 1, where a higher value indicates a greater risk impact on that type of anomaly. For example, a fallen tree may have a higher α_i value than localized overheating. This step ensures that the system can prioritize high-risk events.

[0031] Next, the cumulative number N_i for each type of anomaly is calculated, and the comprehensive impact index I = Σ(α_i × N_i) is obtained by weighting by type. I serves as an indicator of overall risk, representing the sum of the impacts of various anomaly types. For example, if 10 instances of equipment failure (α_i = 0.8) and 5 instances of line aging (α_i = 0.6) are found in a certain area, then I = (0.8 × 10) + (0.6 × 5) = 11.

[0032] Combining distance parameters, the final definition is P = (I / max_I) × (D / max_D), where D is the distance of the current task from the power center, and max_D is the maximum distance. This formula integrates risk and geographical location, increasing the priority of tasks with high-risk anomalies in remote areas. Through this method, resources can be rationally allocated according to urgency, effectively improving response efficiency and security, thereby achieving an intelligent and efficient power inspection system.

[0033] The business intelligent agent in step S4 can specifically refer to unmanned aerial vehicles (UAVs equipped with high-definition visible light, thermal infrared, lidar, etc.) or UAVs carrying other plug-in devices, such as snow removal equipment, pruning equipment, and deterrent equipment. UAVs carrying specific plug-in devices can be dispatched according to task requirements to form an integrated "reconnaissance-processing" intelligent agent. Step S4 involves obtaining the currently available business intelligent agents for performing inspection tasks, i.e., obtaining the number of idle business intelligent agents in real time, facilitating timely planning and allocation of inspection tasks. This can be implemented by connecting to a drone airport (Drone-in-a-Box) management system or a drone cloud platform. The status (idle, busy, charging, faulty), battery level, current location, and the type of equipment mounted on all registered intelligent agents are queried in real time. Based on the task requirements generated in step S3 (such as pruning), idle business intelligent agents with sufficient battery power and the corresponding equipment are automatically selected and notified to prepare for the task.

[0034] In a specific embodiment, the number of business agents M can be obtained first; then the capability level E_i of each agent can be evaluated (e.g., flight time, type of carried sensors, communication stability, etc.); and the priority can be calculated according to the formula, priority = (E_i / max_E) + (remaining battery S_i / max_S), where max_E and max_S are the highest capability and the highest battery, respectively. Select the highest priority agents to participate in the inspection task.

[0035] In step S5, the specific implementation process includes: the central control system sends the task and path information generated in step S3 to the selected drone in step S4 via a 4G / 5G network. Automatic execution scenario: The drone automatically takes off from the airport, flies autonomously along a predetermined route, and achieves centimeter-level precise hovering through RTK positioning. Upon reaching each task point, it automatically adjusts the gimbal to perform tasks such as taking photos, recording videos, and laser scanning, and transmits the data back in real time. For example, the drone automatically performs actions such as flame de-icing, laser pruning, and sonic deterrence. Human-machine collaboration scenario: The high-definition data transmitted by the drone can be streamed to the monitoring center in real time for maintenance personnel to confirm. When necessary, personnel can remotely intervene through the large model's dialogue interface and issue new commands (such as: "Get closer to take a picture"). After completing the task, the drone automatically returns to base and recharges.

[0036] After all data is transmitted back, the large model's generation capability is invoked again to automatically integrate satellite imagery analysis results and UAV close-range inspection data to generate a complete, illustrated inspection analysis report, including hazard confirmation, handling results, and recommended measures. This report is then automatically archived into the production management system, forming a closed loop. Furthermore, in this embodiment, if multiple inspection tasks exist, before controlling the business intelligence agent to proceed along the planned inspection route, the method further includes: determining the priority of the inspection tasks; matching each inspection task with a business intelligence agent in descending order of priority; and traversing all inspection tasks until all inspection tasks are matched with corresponding business intelligence agents. To improve inspection efficiency, the goal is to minimize inspection time. Inspection tasks are matched with business intelligence agents in a many-to-one or one-to-one manner. Depending on the urgency of the task, higher-priority inspection tasks can be prioritized.

[0037] In one specific embodiment, the process involves calculating the total route mileage L based on GPS coordinates; Furthermore, the density of obstacles O_i that may appear along the route is evaluated, and the obstacle index is derived by combining the map information. The route efficiency can be calculated using the following formula: Route efficiency = 1 / (O_i × weight_O + route offset × weight_D) / total_weight, where weight_O and weight_D are the corresponding weight parameters. The most efficient route is selected from all candidate routes for execution.

[0038] Specifically, the total route mileage L can be calculated based on GPS coordinates. L is the actual length of the inspection path, measured in kilometers. Its range is determined by map data and satellite positioning results; a smaller value indicates a shorter path, which helps reduce energy consumption. Secondly, the obstacle density O_i along the route is assessed. O_i is the number of obstacles that may be encountered per unit length, such as trees, buildings, or natural barriers, ranging from 0 to 1. A lower value indicates a smoother path. An obstacle index is then derived by combining map information. This index is a weighted cumulative sum of O_i and is used to measure path safety. Next, the path efficiency is calculated using the formula: "Route efficiency = 1 / ((O_i × weight_O + route offset × weight_D) / total_weight)," where weight_O and weight_D are weight parameters, representing the degree of attention paid to obstacle density and route offset, respectively, typically between 0.5 and 1. total_weight is the sum of all weights, i.e., the sum of all weight_O and weight_D for the path. Higher route efficiency results in a better path selection. Finally, the most efficient route is selected from all candidate routes to ensure efficient and safe inspection operations. For example, in a mountainous power grid inspection, the agent compared two routes, finding that one, although shorter, had dense vegetation and many high-risk obstacles, while the other, though a longer route, had smoother roads and less offset. The latter was ultimately chosen, achieving higher safety and work efficiency. This method improves the automation level of inspections, enhances adaptability to complex geographical environments, reduces the need for manual intervention, and improves the operation and maintenance efficiency of the power system.

[0039] Furthermore, in this embodiment of the application, the method further includes: analyzing the equipment requirements of the inspection task for external plug-in devices (e.g., whether there is snow removal, pruning, or driving away); establishing business intelligent agents for different inspection tasks, and adding the necessary external plug-in devices to different business intelligent agents.

[0040] Example 2

[0041] Example 1 uses an automatic generation method for inspection tasks, while this example provides a user-defined method for generating inspection tasks, specifically: A1. Obtain user instructions; the user instructions may include text instructions or voice instructions. A2. Parse user instructions to determine user inspection requirements; A3. Develop inspection tasks based on user inspection needs.

[0042] Specifically, in step A1, the large model's multi-turn dialogue capabilities can be used to obtain user instructions. For example, the user issues an initial instruction, such as a vague voice message: "Help me check the poles and tree obstructions on that section of the line in the XX mountain area after last week's heavy rain." The Automatic Speech Recognition (ASR) module converts the speech into text. The large model takes over the dialogue; it can identify missing information in the instruction (such as the specific line number, more precise section) and proactively ask follow-up questions to clarify the request. For example, it might reply: "Okay, we are creating an inspection task for you. Please confirm that you are referring to the section between towers A12 and A27 of the '500kV Pingyong Line'? Furthermore, do you want to focus on checking for pole foundation landslides, tower tilting, or fallen trees in the line corridor?" Through one or more such interactions, the large model gradually guides the user to supplement and improve the information, ultimately forming a clear and unambiguous task instruction.

[0043] In step A2, user instructions are parsed to determine user inspection needs, including: writing inspection needs based on pre-established prompt word templates; obtaining prompt words from user instructions; and identifying user inspection needs based on prompt words and contextual information. Specifically, the system pre-sets prompt word templates for different inspection scenarios. The purpose of these templates is to "guide" and "frame" the thinking direction of the large model, ensuring the structured nature of its output. The large model combines user instructions (as dynamic prompt words) with the aforementioned static prompt word templates. Based on powerful contextual understanding capabilities, it accurately extracts key information from the dialogue history and populates it into the fields specified in the template.

[0044] In step A3, inspection tasks are formulated based on user inspection requirements. Specifically, this includes: breaking down the task objective (which may have multiple anomalies of varying types) into several sub-tasks (i.e., multiple inspection tasks); analyzing the dependencies and priorities between each sub-task; and orchestrating the tasks based on their priorities and dependencies, matching the sub-tasks with multiple business intelligence agents and distributing them to each agent. The sub-tasks are provided by a pre-defined list, which records the task name, description, implementation method, and relationships between sub-tasks. The multiple business intelligence agents are provided by a business intelligence agent list, which records specific information about each agent, including their API calls and functional descriptions. Figure 2 As shown.

[0045] Finally, a detailed and executable overall task plan is generated, which includes: a task list: a detailed description, parameters, and priority of each subtask; an allocation scheme: which agent each subtask is assigned to; and an execution sequence: the optimal execution route map that takes into account dependencies and geographical location.

[0046] The steps for prioritizing and arranging the above-mentioned integrated tasks, matching several sub-tasks with multiple business intelligence agents, and distributing them to each business intelligence agent are as follows: Based on the pre-defined list of subtasks, the priority and dependencies between subtasks are confirmed to obtain the task orchestration process; Match the subtask preset list and the business intelligence agent list, traverse all subtasks until all subtasks have been matched with the corresponding business intelligence agents.

[0047] As a preferred embodiment, the steps for completing the corresponding task based on a pre-established business intelligence agent include: Establish a business intelligence agent based on user business needs; Each business intelligence agent is bound to one or more subtasks; After receiving the assigned subtask, the business intelligence agent executes it according to the task orchestration process; The large model summarizes the execution results of the business intelligence agents, uses their understanding capabilities to summarize the results, and then outputs them.

[0048] Example 3

[0049] like Figure 3 As shown, this embodiment provides a power line inspection system based on satellite remote sensing and a large model. The system includes: a satellite module 1, used to acquire satellite remote sensing images of a target area; a large model module 2, used to determine whether there are abnormal features in the target area based on the satellite remote sensing images; when abnormal features are present, it is also used to determine the inspection task and plan the inspection route based on the number and location of the abnormal features; a status monitoring module 3, used to monitor the working status of each business agent and acquire the business agents currently available to perform inspection tasks; and a business agent 4, used to perform inspection tasks by traveling along the planned inspection route.

[0050] Furthermore, the system also includes a visualization module, which is based on a WebGIS engine and uses an automatic publishing method to visualize the spatial data results output by the large model, making it convenient for users to view the inspection results and make manual corrections.

[0051] The system provided in this embodiment can be used to execute the methods described in Embodiments 1 and 2, as detailed in Embodiments 1 and 2, which will not be repeated here.

[0052] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0053] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0054] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0055] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0056] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0057] Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0058] Computer-readable media include both permanent and non-permanent, removable and non-removable media, which can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0059] It should also be noted that 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 process, method, article, or apparatus. Unless otherwise specified, 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 that element.

[0060] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A power line inspection method based on satellite remote sensing and a large model, characterized in that, The method includes: Acquire satellite remote sensing images of the target area; Based on the satellite remote sensing imagery, determine whether there are any abnormal features within the target area; If they exist, the inspection task and route are determined based on the number and location of the abnormal features, including: counting the total number of abnormal points N; calculating the priority factor P = (D / max_D) * (N / max_N) based on the distance D between the abnormal point and the power station, where max_D and max_N are the maximum distance and the maximum number of abnormalities, respectively; and generating the inspection order based on the P value. Obtain the business intelligence agents currently available to perform inspection tasks; The control business intelligence agent moves along the planned inspection route to complete the inspection task.

2. The power line inspection method based on satellite remote sensing and a large model according to claim 1, characterized in that, If there are multiple inspection tasks, the method further includes the following before the control agent proceeds according to the planned inspection route: Determine the priority of inspection tasks; The business intelligence agents are matched to each inspection task in descending order of priority. Iterate through all inspection tasks until each inspection task is matched with the corresponding business intelligence agent.

3. The power line inspection method based on satellite remote sensing and a large model according to claim 1, characterized in that, If there are multiple inspection tasks, the method further includes the following before the control agent proceeds according to the planned inspection route: With the goal of minimizing inspection time, inspection tasks are matched with business intelligence agents in a many-to-one or one-to-one manner.

4. The power line inspection method based on satellite remote sensing and a large model according to claim 1, characterized in that, The method further includes: Analyze the equipment requirements of external plug-in devices for inspection tasks; Establish separate business intelligence agents for different inspection tasks, and add necessary external plug-in devices to each business intelligence agent.

5. The power line inspection method based on satellite remote sensing and a large model according to claim 1, characterized in that, The method further includes: Obtain user instructions; the user instructions may include text instructions or voice instructions. Parse user commands to determine user inspection needs; Inspection tasks are formulated based on user inspection requirements.

6. The power line inspection method based on satellite remote sensing and a large model according to claim 5, characterized in that, The process of parsing user commands and determining user inspection needs includes: Write inspection requirements based on pre-established prompt word templates; Obtain prompt words from user instructions, and identify user inspection needs based on prompt words and context.

7. The power line inspection method based on satellite remote sensing and a large model according to claim 1, characterized in that, The business intelligence agent includes unmanned cameras.

8. A power line inspection system based on satellite remote sensing and a large model, characterized in that, The system includes: The satellite module is used to acquire satellite remote sensing images of the target area; The large model module is used to determine whether there are abnormal features in the target area based on the satellite remote sensing image. When abnormal features are found, it is also used to determine the inspection task and plan the inspection route based on the number and location of the abnormal features. Specifically, if abnormal features are found, the inspection task and route planning are determined based on the number and location of the abnormal features, including: counting the total number of abnormal points N; calculating the priority factor P = (D / max_D) * (N / max_N) based on the distance D between the abnormal points and the power station, where max_D and max_N are the maximum distance and the maximum number of abnormal points, respectively; and generating the inspection order based on the P value. The status monitoring module is used to monitor the working status of each business intelligence agent and to obtain the business intelligence agents that can currently be used to perform inspection tasks. The business intelligence agent is used to perform inspection tasks by following the planned inspection route.

9. A processor, characterized in that, It is configured to perform the power line inspection method based on satellite remote sensing and large model as described in any one of claims 1 to 7.

10. A machine-readable storage medium storing instructions thereon, characterized in that, When executed by a processor, the instruction causes the processor to be configured to perform the power line inspection method based on satellite remote sensing and large models as described in any one of claims 1 to 7.

Images