A dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations

Abstract

The present application discloses a dynamic planning method for intelligent line selection of power transmission lines based on segmentation considerations, which relates to the field of power engineering. The method comprises: generating a potential path point sequence according to starting point information, end point information and environmental data through a large language model, wherein the starting point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located; performing a two-dimensional path search on the potential path point sequence through a target algorithm on a terrain feature layer to generate an initial path of the overhead transmission line; performing three-dimensional elevation optimization on the path nodes in the initial path according to digital elevation data to obtain a final path of the overhead transmission line; adopting the above scheme solves the problem that the overhead transmission line line selection planning process has many influencing factors and they restrict each other, manual line selection is time-consuming and labor-intensive, planning efficiency is low and line plan quality is poor.

Description

Technical Field

[0001] The present application relates to the field of electric power engineering, and in particular to a dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations. Background Art

[0002] Route planning for overhead transmission lines is a critical step in power engineering design, requiring minimal cost and environmental impact while meeting safety regulations. Traditionally, transmission line selection relies heavily on manual experience: designers draw on data such as topographic maps and satellite imagery to plan a corridor between the starting and ending points, avoiding obstacles. Route planning requires consideration of various ground features (such as roads, rivers, buildings, and vegetation), as well as terrain elevation (undulating terrain). Routes must minimize length, avoid crossing restricted areas (such as densely built-up areas and nature reserves), and minimize intersections with linear infrastructure such as roads and railways. Due to the numerous and interdependent factors, manual route selection often requires repeated weighing of multiple options, which is time-consuming and labor-intensive.

[0003] In related technologies, there is no effective solution to the problems that the overhead transmission line selection and planning process is affected by many factors and they restrict each other, manual line selection is time-consuming and labor-intensive, the planning efficiency is low, and the line plan quality is poor. Summary of the Invention

[0004] The embodiment of the present application provides a dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations, so as to at least solve the problems in related technologies, that is, the manual line selection is time-consuming and labor-intensive, the planning efficiency is low, and the line plan quality is poor due to the numerous factors affecting the line selection planning process of overhead transmission lines and their mutual constraints.

[0005] According to one aspect of an embodiment of the present application, a dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations is provided, comprising: generating a sequence of potential path points according to starting point information, end point information and environmental data through a large language model, wherein the starting point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located; performing a two-dimensional path search on the sequence of potential path points through a target algorithm on a feature layer to generate an initial path of the overhead transmission line; and performing three-dimensional elevation optimization on the path nodes in the initial path according to digital elevation data to obtain a final path of the overhead transmission line.

[0006] In an exemplary embodiment, a two-dimensional path search is performed on the potential path point sequence through a target algorithm on a feature layer to generate an initial path of the overhead transmission line, including: initializing an exploration graph, wherein the starting point information is the root node of the exploration graph; determining multiple candidate points in the potential path point sequence whose distance from the root node is less than a first preset threshold as an initial priority sequence, wherein the multiple candidate points are used to indicate multiple path directions of the initial path; cyclically expanding the exploration graph through a potential point priority mechanism and / or a random node sampling method until a node in the exploration graph is connected to a path node corresponding to the end point information or the number of cyclic extensions reaches a preset number of iterations; extracting a path sequence from the exploration graph based on the root node and the path node corresponding to the end point information, and generating the initial path based on the path sequence.

[0007] In an exemplary embodiment, the exploration graph is cyclically expanded through a potential point priority mechanism and / or a random node sampling method, including: when it is determined that there is a first path node in the potential path point sequence that has not been added to the exploration graph, determining a second path node with the highest priority among the first path nodes; determining whether the subpath between the newly added third path node in the exploration graph and the second path node complies with the route planning rules, wherein the route planning rules are used to indicate that the initial path does not overlap with the area where the obstacle is located, and the subpath is a straight path; when it is determined that the route planning rules are complied with, the second path node is added to the exploration graph; when it is determined that the first path node does not exist or is not complied with, a fourth path node is randomly selected from the first sub-area from the third path node to the path node corresponding to the end point information through the random node sampling method, and when it is determined that the subpath between the third path node and the fourth path node complies with the route planning rules, the fourth path node is added to the exploration graph, wherein the target area includes the first sub-area.

[0008] In an exemplary embodiment, the method further includes: when the number of failures in expanding the exploration graph through the potential point priority mechanism reaches a second preset threshold, increasing the sampling frequency of the random node sampling method.

[0009] In an exemplary embodiment, before performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain the final path of the overhead transmission line, the method further includes: identifying non-critical nodes on multiple path nodes in the initial path to obtain multiple non-critical nodes, wherein the sub-path where the non-critical nodes are located does not overlap with the area where the obstacle is located, and the sub-path is a straight path; deleting the multiple non-critical nodes from the initial path to obtain a simplified path; and performing three-dimensional elevation optimization on the path nodes in the simplified path according to the digital elevation data to obtain the final path.

[0010] In an exemplary embodiment, three-dimensional elevation optimization is performed on the path nodes in the simplified path based on the digital elevation data to obtain the final path, including: determining the altitudes of multiple fifth path nodes in the simplified path based on the digital elevation data; for a sixth path node among the multiple fifth path nodes, searching for a local terrain high point in the second sub-area where the sixth path node is located, and updating the sixth path node to the local terrain high point to obtain multiple updated fifth path nodes, wherein the local terrain high point is a legal construction area; and determining the final path based on the updated multiple fifth path nodes.

[0011] In an exemplary embodiment, determining the final path based on the updated plurality of fifth path nodes includes: sequentially determining horizontal distances between two adjacent fifth path nodes in the updated plurality of fifth path nodes, and determining whether any of the plurality of horizontal distances has a target horizontal distance greater than a third preset threshold; if it is determined that any target horizontal distance is greater than the third preset threshold, adding one or more intermediate path nodes between the two fifth path nodes corresponding to the target horizontal distances to obtain a plurality of seventh path nodes, such that the horizontal distance between any two adjacent seventh path nodes is less than or equal to the third preset threshold; sequentially determining slope values ​​of a pole-tower connection corresponding to two adjacent seventh path nodes in the plurality of seventh path nodes, and determining whether any of the plurality of slope values ​​has a target slope value greater than a fourth preset threshold; if it is determined that any target slope value is greater than the fourth preset threshold, adjusting the heights of the two seventh path nodes corresponding to the target slope values ​​to obtain a plurality of eighth path nodes, such that the slope values ​​corresponding to any two adjacent eighth path nodes are less than or equal to the fourth preset threshold; and determining the final path based on the plurality of eighth path nodes.

[0012] According to another aspect of an embodiment of the present application, a multi-stage route planning device based on large model recommendation and terrain optimization is also provided, including: a generation module for generating a potential path point sequence based on starting point information, end point information and environmental data through a large language model, wherein the starting point information and the end point information are used to build an overhead transmission line, and the environmental data is used to indicate the environmental information of the target area where the overhead transmission line is located; a search module for performing a two-dimensional path search on the potential path point sequence through a target algorithm on a terrain layer to generate an initial path of the overhead transmission line; an optimization module for performing three-dimensional elevation optimization on the path nodes in the initial path according to digital elevation data to obtain the final path of the overhead transmission line.

[0013] According to another aspect of the embodiment of the present application, a computer-readable storage medium is also provided, in which a computer program is stored, wherein the computer program is configured to execute the above-mentioned dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations when running.

[0014] According to another aspect of an embodiment of the present application, an electronic device is also provided, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the above-mentioned dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations through the computer program.

[0015] According to another aspect of the embodiments of the present application, a computer program product is provided, including a computer program, which implements the steps of the method described in each embodiment of the present application when executed by a processor.

[0016] In an embodiment of the present application, a dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations is proposed. The method first processes the starting point information, end point information and environmental data through a large language model to generate a sequence of potential path points, wherein the starting point information and end point information are respectively used to indicate the starting point and end point of the overhead transmission line to be constructed, and the environmental data are used to indicate the environmental information in the target area where the overhead transmission line is located; then, a target algorithm is used to recommend potential path points and quickly grow the path to obtain an initial path; finally, the generated initial path is fine-tuned in elevation and supplemented with nodes, thereby completing the fine optimization of the line to obtain the final path; the above scheme is adopted to combine the environmental understanding ability of the large language model and the path search algorithm, and the digital elevation model is introduced in stages for optimization to achieve fully automatic line selection planning for overhead lines; through this technical scheme, the line selection efficiency can be significantly improved, the dependence on manual experience can be reduced, and a safer, more reliable and economically reasonable line plan can be generated; thereby solving the problem in the related art that the line selection planning process of overhead transmission lines is affected by many factors and they restrict each other, manual line selection is time-consuming and labor-intensive, the planning efficiency is low and the line plan quality is poor. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0018] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0019] Figure 1 is a flow chart of an optional dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations according to an embodiment of the present application;

[0020] Figure 2 This is a flow chart of an optional dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations according to an embodiment of the present application;

[0021] Figure 3 This is a structural block diagram of an optional multi-stage route planning device based on large model recommendation and terrain optimization according to an embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0023] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0024] In related technologies, the development of geographic information systems (GIS) has led to the emergence of GIS-based automatic transmission line selection solutions. For example, one method converts environmental constraints into grid data and applies a shortest path algorithm to search for the optimal path. Some research utilizes intelligent optimization algorithms to assist in line selection: for example, combining the ant colony algorithm with GIS, using the fuzzy analytic hierarchy process on the ArcGIS platform to fuse multiple geographic layers to generate a comprehensive cost layer, and then searching for the path with the lowest overall cost. Other research has improved the classic Dijkstra and A* algorithms, reducing invalid operations by setting efficient priority search areas, thereby improving the efficiency of path search in large-scale environments.

[0025] Improved RRT (Rapidly-exploring Random Tree) path planning methods (such as RRT* and Informed-RRT*) can achieve path optimization through sampling optimization, but they still require extensive trial sampling across a vast space, consuming significant computational resources. Furthermore, their treatment of typical features such as buildings, roads, and rivers is relatively crude, which can easily lead to paths crossing obstacles or nodes falling within unbuildable areas. Some deep learning-based methods are solely used for classification and recognition and do not participate in the actual path generation process.

[0026] In recent years, the rise of Large Language Models (LLMs) has provided new insights into path planning. Some research has attempted to combine the global reasoning capabilities of LLMs with traditional path planning algorithms. For example, LLMs are used in conjunction with the A* algorithm to plan paths, leveraging the LLM's macroscopic understanding of the environment to guide the algorithm's search direction. These approaches have significantly improved the efficiency and intelligence of path planning.

[0027] Most existing transmission line planning tools focus on two-dimensional route optimization. After generating a preliminary route, manual 3D elevation verification and tower site selection are performed. Typically, towers (line support points) must be positioned away from buildings and located as high as possible to maximize conductor clearance. Furthermore, tower spacing should be kept within a reasonable range to control conductor sag and mechanical strength. However, manual route selection is time-consuming and labor-intensive, and relies heavily on experience.

[0028] In order to solve the technical problems existing in the related art, this embodiment provides a dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations. Figure 1 This is a flowchart of an optional segment-based dynamic planning method for intelligent line selection of transmission lines according to an embodiment of the present application. The process includes the following steps S102-S106:

[0029] Step S102: generating a potential pathpoint sequence based on the starting point information, the ending point information, and the environmental data using a large language model, wherein the starting point information and the ending point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located;

[0030] Step S104, performing a two-dimensional path search on the potential path point sequence on the feature layer using a target algorithm to generate an initial path of the overhead transmission line;

[0031] Step S106 , performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line.

[0032] Through the above steps, a dynamic planning method for intelligent transmission line selection based on segmentation considerations is proposed. The method first processes the starting point information, end point information and environmental data through a large language model to generate a sequence of potential path points, where the starting point information and end point information are used to indicate the starting point and end point of the overhead transmission line to be constructed, respectively, and the environmental data is used to indicate the environmental information in the target area where the overhead transmission line is located. Then, a target algorithm is used to recommend potential path points and quickly grow the path to obtain an initial path. Finally, the generated initial path is fine-tuned in elevation and supplemented with nodes to complete the fine optimization of the line and obtain the final path. The above scheme combines the environmental understanding capabilities of the large language model and the path search algorithm, and introduces the digital elevation model for optimization in stages to achieve fully automatic line selection planning for overhead lines. Through this technical scheme, the line selection efficiency can be significantly improved, the dependence on manual experience can be reduced, and a safer, more reliable and economically reasonable line plan can be generated. This solves the problem in related technologies that the overhead transmission line line selection planning process is time-consuming and labor-intensive, with low planning efficiency and poor line plan quality due to the numerous influencing factors and mutual constraints in the process.

[0033] In an exemplary embodiment, a two-dimensional path search is performed on the potential path point sequence through a target algorithm on a feature layer to generate an initial path of the overhead transmission line, including: initializing an exploration graph, wherein the starting point information is the root node of the exploration graph; determining multiple candidate points in the potential path point sequence whose distance from the root node is less than a first preset threshold as an initial priority sequence, wherein the multiple candidate points are used to indicate multiple path directions of the initial path; cyclically expanding the exploration graph through a potential point priority mechanism and / or a random node sampling method until a node in the exploration graph is connected to a path node corresponding to the end point information or the number of cyclic extensions reaches a preset number of iterations; extracting a path sequence from the exploration graph based on the root node and the path node corresponding to the end point information, and generating the initial path based on the path sequence.

[0034] This example describes a method for initial path planning for overhead transmission lines using a large-scale model recommendation and a rapidly-exploring random tree (RRT) algorithm. This method combines intelligent recommendation with random exploration to efficiently generate a short and direct preliminary path that avoids obstacles.

[0035] 1. Initialize the exploration graph:

[0036] The construction of the exploration graph begins with the definition of a root node, which is the starting point of the transmission line. The starting point information includes the starting point coordinates (X_start, Y_start) and possible additional attributes such as altitude and terrain type.

[0037] Construct an empty exploration graph structure, insert the starting point node into the tree as the root node of the tree, and mark it as RRT_tree ={root}.

[0038] 2. Generate potential path point sequence:

[0039] The DeepSeek language model is used to generate a series of potential path points based on the starting point, end point, and environmental information. These points are intelligently recommended based on factors such as topography and obstacle distribution to indicate the most likely path direction.

[0040] The potential path point sequence preliminarily determines the search priority and direction through its relative position relationship with the starting point. For example, the distance and orientation information of points P1, P2, P3, etc. in the sequence from the starting point will guide the expansion direction of the exploration graph.

[0041] 3. Potential point priority mechanism and random node sampling:

[0042] Priority sequence initialization: A number of candidate points whose distance from the root node is less than a preset threshold (e.g., 100 meters) are selected from the potential pathpoint sequence to form an initial priority sequence. These points will be prioritized during the search process to guide the growth of the exploration graph.

[0043] Loop extension: Before the exploration graph reaches the end, or the number of iterations does not exceed the preset value N_max, the algorithm performs loop extension:

[0044] Priority expansion: According to the priority sequence, select the nearest candidate point P_near as the expansion target. Find the node N_near closest to P_near in the tree, try to sample a new node P_new from N_near to P_near, and check whether it intersects with the obstacle.

[0045] Collision detection: If the new node P_new and its connection to N_near do not intersect with obstacles such as buildings and rivers, P_new is added to the exploration graph and its path to the root node is updated.

[0046] Random expansion: When the potential point priority mechanism is blocked from expansion (e.g., when P_new intersects an obstacle), the algorithm switches to random mode, randomly sampling a node R_rand in free space. The tree then finds the node N_rand closest to R_rand, attempts to generate a new node R_new, and checks for collisions.

[0047] Priority update: Whenever a new node is successfully added, the priority order in the potential path point sequence is dynamically updated based on its location information to ensure that subsequent expansion is closer to the target path.

[0048] 4. Path generation and extraction:

[0049] End point determination: If the node generated during the cyclic expansion process is within a preset connection distance (e.g., 10 meters) from the node corresponding to the end point information, the path is determined to be complete and the expansion is terminated.

[0050] Path sequence extraction: Once the exploration graph reaches the end point, start from the root node and trace the shortest path to the end point in reverse to extract the path sequence, which is a series of connected tower node locations.

[0051] Initial path generation: Based on the extracted path sequence, a preliminary overhead transmission line path is generated, that is, the line layout represented on a two-dimensional plane.

[0052] This implementation utilizes the potential path points recommended by a large model in conjunction with the exploration graph, cleverly integrating intelligent guidance with random exploration to achieve efficient and intelligent preliminary path planning for overhead transmission lines. By prioritizing recommended points, the inefficiency and path deviations that can result from blind random searches are avoided. Random node sampling ensures comprehensiveness and flexibility in the search, ensuring that feasible paths can be found even in complex environments. This approach significantly improves the automation and quality of path planning, providing a solid foundation for subsequent elevation optimization and line engineering design.

[0053] Optionally, the exploration graph is cyclically expanded through a potential point priority mechanism and / or a random node sampling method, including: when it is determined that there is a first path node in the potential path point sequence that has not been added to the exploration graph, determining the second path node with the highest priority among the first path nodes; determining whether the sub-path between the newly added third path node in the exploration graph and the second path node complies with the route planning rules, wherein the route planning rules are used to indicate that the initial path does not overlap with the area where the obstacle is located, and the sub-path is a straight line path; when it is determined that the route planning rules are complied with, the second path node is added to the exploration graph; when it is determined that the first path node does not exist or is not complied with, a fourth path node is randomly selected from the first sub-area from the third path node to the path node corresponding to the end point information through the random node sampling method, and when it is determined that the sub-path between the third path node and the fourth path node complies with the route planning rules, the fourth path node is added to the exploration graph, wherein the target area includes the first sub-area.

[0054] In this example, we explain how to cyclically expand a Rapid Exploration Random Tree (RRT) using a potential point prioritization mechanism and random node sampling to efficiently generate a preliminary path that adheres to route planning rules. This process focuses on flexibly and intelligently expanding the exploration graph to ensure that the path avoids obstacles while rapidly advancing toward the destination.

[0055] 1. Prioritize expansion of potential points:

[0056] Priority check: First, check whether there are any first path nodes in the potential path point sequence that have not been added to the exploration graph. These nodes are previously recommended by the large model and are used to guide the direction of path search.

[0057] Determine the highest priority point: If these points exist, the algorithm will select the one with the highest priority (the second path node), which is determined based on factors such as distance from the starting point, relative position between the recommended points, and path planning rules.

[0058] Subpath generation and verification: The algorithm attempts to generate a straight path (subpath) between the nearest third path node (the most recently added node in the exploration graph) and the second path node, and checks whether this subpath overlaps with any obstacle areas, including roads, rivers, buildings, etc.

[0059] Subpaths are added to the exploration graph: If the subpath meets the route planning rules, that is, it does not overlap with obstacles, the second path node will be added to the exploration graph, expanding the scope of the tree and updating the best path from the starting point to the current point.

[0060] 2. Random node sampling expansion:

[0061] When potential points are exhausted or do not meet the rules: If all points in the potential path point sequence have been added to the exploration graph, or all subpaths between potential points and nodes in the exploration graph do not meet the route planning rules, the algorithm switches to random node sampling mode.

[0062] Target area setting: Define a first sub-area from the third path node (the nearest node in the exploration graph) to the node corresponding to the end point information. This area is the key area searched by the current algorithm.

[0063] Randomly select the fourth path node: randomly select a point (the fourth path node) in the first sub-area, and check whether the sub-path from the third path node to the random point complies with the route planning rules.

[0064] Subpath verification and joining: If this randomly generated subpath passes the obstacle detection, that is, it does not overlap with any obstacle area, then the fourth path node will be added to the exploration graph and the tree expansion process will continue.

[0065] 3. Loop conditions and termination conditions:

[0066] Loop condition: As long as the exploration graph has not reached the node corresponding to the end information and the number of loops does not exceed the preset maximum number of iterations, the loop expansion will continue.

[0067] Termination condition: Once the node in the exploration graph is within the preset connection distance from the node corresponding to the end point information, or the number of loops reaches the preset maximum value, the loop stops and the path planning is considered complete.

[0068] Summary: By combining a potential point prioritization mechanism with a random node sampling method, the exploration graph expansion process achieves a balance between intelligence and flexibility. Potential point prioritization accelerates path progression toward the endpoint and reduces blind searches; random sampling ensures the algorithm's adaptability to complex environments and avoids the local optimality trap of path planning. This expansion strategy enables efficient growth of the exploration graph, ultimately generating paths that avoid all obstacles while remaining as close to a straight line as possible, shortening line distances and improving the intelligence and efficiency of overhead transmission line planning.

[0069] Optionally, the method further includes: when the number of failures in expanding the exploration graph through the potential point priority mechanism reaches a second preset threshold, increasing the sampling frequency of the random node sampling method.

[0070] In this example, we introduce how to dynamically adjust the algorithm strategy and increase the frequency of random node sampling when encountering difficulties in expanding the exploration graph based on the potential point priority mechanism to overcome the problem of local search limitations and ensure the continuous progress and ultimate success of path planning.

[0071] 1. Limitations of the potential point priority mechanism:

[0072] In the initial path planning of overhead transmission lines, although a potential point prioritization mechanism is used to guide the expansion of the exploration graph, due to the complexity of the terrain and the uncertainty of obstacle distribution, it is possible that multiple potential point-based expansion attempts will fail. This failure may occur because the potential point is located near an obstacle, the terrain is too rugged to connect directly, or other planning rules restrict the expansion.

[0073] 2. Failure threshold:

[0074] To prevent the algorithm from falling into an infinite loop or being overly constrained by the potential point mechanism, this embodiment sets a second preset threshold to monitor the number of consecutive failures. This threshold can be an empirical value derived from experimental testing, such as 10 consecutive unsuccessful attempts at potential point-based expansion.

[0075] 3. Increase random sampling frequency:

[0076] When the number of failures in exploring the graph expansion through the potential point priority mechanism reaches a second preset threshold, the algorithm automatically adjusts the strategy and increases the frequency of using the random node sampling method.

[0077] This means that in subsequent loop expansions, the algorithm will increasingly employ random node sampling, rather than relying solely on guidance from potential points. Random sampling allows for the exploration of potential paths in a wider free space, increasing the diversity and possibilities of the search. This is particularly true in environments with complex terrain and dense obstacles, where it helps the algorithm escape local optima and discover new feasible paths.

[0078] 4. Restore potential point priority mechanism:

[0079] When random node sampling successfully generates a new node and adds it to the RFT tree, the algorithm will re-evaluate the current exploration graph structure and environmental conditions. If the expansion of the potential point priority mechanism becomes feasible or effective again, the algorithm will restore the original proportion of potential point priority expansion combined with random node sampling.

[0080] In this way, the algorithm can adopt a more flexible search strategy when the expansion of potential points is blocked. Once the obstacles are cleared or the path conditions improve, the intelligently recommended potential points can be reused for targeted search to achieve efficient and intelligent path planning.

[0081] In summary, this embodiment enhances the exploration graph's path planning capabilities in complex environments by dynamically adjusting the frequency of using the potential point prioritization mechanism and random node sampling. This allows the algorithm to more robustly address various challenges and ensures that, even under adverse conditions, a preliminary overhead transmission line path that satisfies all planning rules can ultimately be generated. This approach not only improves path planning efficiency but also enhances the algorithm's robustness and generalization capabilities, making it a key approach to advancing automation in the field of intelligent path planning.

[0082] Optionally, before performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain the final path of the overhead transmission line, the method further includes: identifying non-critical nodes on multiple path nodes in the initial path to obtain multiple non-critical nodes, wherein the sub-path where the non-critical nodes are located does not overlap with the area where the obstacle is located, and the sub-path is a straight line path; deleting the multiple non-critical nodes from the initial path to obtain a simplified path; and performing three-dimensional elevation optimization on the path nodes in the simplified path according to the digital elevation data to obtain the final path.

[0083] In this example, we explored how to simplify the initial path of an overhead transmission line before applying digital elevation data for 3D elevation optimization to eliminate non-critical nodes, reduce algorithm complexity, and ensure that the simplified path still meets safety and regulatory requirements.

[0084] 1. Identification of non-critical nodes:

[0085] After the initial path is obtained, the path may contain a large number of path nodes. Some of these nodes are path turning points and play a key role in obstacle avoidance and path direction; while other nodes are merely by-products of the path growth process and have little or no impact on the final path.

[0086] Path Analysis and Subpath Checking: The algorithm examines each straight subpath (segment) formed between each pair of adjacent nodes on the initial path and evaluates whether this subpath overlaps with any obstruction areas (such as buildings, rivers, roads, etc.). If the subpath does not overlap with obstacles and the straight segment of the path can pass directly through free space, the path node on this straight segment is considered a non-critical node.

[0087] 2. Non-critical node marking: All path nodes that meet the above conditions will be marked as non-critical nodes. These nodes will be deleted before subsequent elevation optimization.

[0088] Path simplification:

[0089] Node deletion and path reconstruction: After identifying all non-critical nodes, the algorithm will delete them from the initial path, retaining only the critical nodes (i.e., path inflection points), as well as the starting and ending points.

[0090] Path reconstruction: After deleting non-critical nodes, a straight path is re-formed between adjacent critical nodes. This simplified path only includes nodes that have a decisive impact on the final route layout, greatly reducing the computational complexity of subsequent elevation optimization.

[0091] 3. 3D elevation optimization:

[0092] DEM data introduction: After obtaining the simplified path, digital elevation model (DEM) data is used to optimize the elevation of each node along the path. The goal is to locate each tower node in a relatively flat and appropriately elevated position to meet the line's ground clearance requirements and installation specifications, and avoid engineering problems caused by lines that are too low or too high.

[0093] Node fine-tuning and insertion: The algorithm will fine-tune the elevation of each key node of the simplified path so that it falls on relatively high ground. At the same time, it will insert intermediate points when necessary to control the spacing between towers and the slope of the path, ensuring that the line is both safe and economical in three-dimensional space.

[0094] 4. Final path generation:

[0095] The simplified path after elevation optimization is the final path for the overhead transmission line. This path not only avoids all obstacles in the two-dimensional plane but also fully considers terrain factors in three-dimensional space, making the line layout more reasonable and more in line with actual engineering requirements and safety regulations.

[0096] Summary: Through the path simplification and three-dimensional elevation optimization process described in the above embodiments, the present invention provides an efficient and accurate method for overhead transmission line path planning. Path simplification can effectively reduce the consumption of computing resources, while three-dimensional elevation optimization ensures the safety and economy of the line during actual construction. The final path is the optimal solution after multiple optimizations of intelligent recommendation, obstacle avoidance, and terrain adaptability. This method not only improves the efficiency and quality of automated line selection and planning, but also reduces reliance on manual adjustments, providing strong support for the rapid and intelligent design of overhead transmission lines.

[0097] In an exemplary embodiment, three-dimensional elevation optimization is performed on the path nodes in the simplified path based on the digital elevation data to obtain the final path, including: determining the altitudes of multiple fifth path nodes in the simplified path based on the digital elevation data; for a sixth path node among the multiple fifth path nodes, searching for a local terrain high point in the second sub-area where the sixth path node is located, and updating the sixth path node to the local terrain high point to obtain multiple updated fifth path nodes, wherein the local terrain high point is a legal construction area; and determining the final path based on the updated multiple fifth path nodes.

[0098] In this example, we detailed how to use Digital Elevation Model (DEM) data to optimize the three-dimensional elevation of nodes along an overhead transmission line path, ultimately determining a final line path that fully considers the terrain. This process ensures that the line layout is not only rational in plan but also complies with engineering specifications in the vertical dimension, improving the safety and cost-effectiveness of the line layout.

[0099] 1. Determine the altitude of the path node:

[0100] Obtain the elevation information for all fifth-path nodes in the simplified path (the path after removing non-critical nodes). This step is completed by querying the true elevation values ​​corresponding to the coordinates of each node using DEM data, laying the foundation for subsequent elevation optimization.

[0101] 2. Local terrain high point search and node update:

[0102] Target node selection: Select a node from the fifth path node as the sixth path node. This node is usually the key tower location point that requires elevation optimization.

[0103] Second Sub-Area Definition: Define a second sub-area around the sixth path node. The size and shape of this sub-area can be tailored based on terrain complexity and tower layout requirements. This sub-area is used to search for the optimal local terrain high point. For example, the sub-area can be a circular area centered on the sixth path node with a fixed radius (e.g., 100 meters).

[0104] Searching for local terrain high points: Within the defined second sub-area, the algorithm uses DEM data to search for a point (local terrain high point) with the highest elevation that falls within the legal construction area. This involves analyzing terrain information to determine which areas are suitable for tower construction, such as avoiding swamps, cliffs, or protected areas.

[0105] Node position update: After finding a suitable local terrain high point, update the position of the sixth path node to this high point, that is, raise the node in the vertical direction to obtain better terrain adaptability and higher wire clearance from the ground.

[0106] 3. Elevation optimization cycle:

[0107] Repeat the above steps for each fifth (sixth) path node in the simplified path to ensure that all tower locations have been properly optimized. The updated path should be continuously checked for overall compliance during the optimization process, including but not limited to tower spacing restrictions and slope requirements.

[0108] Termination condition: When all nodes have completed elevation optimization or the preset upper limit of optimization iterations has been reached, the algorithm stops executing.

[0109] 4. Determine the final path:

[0110] After optimizing the elevation of all fifth-path nodes, the algorithm reconstructs the entire path based on the updated node positions, forming the final overhead transmission line path. This path not only avoids obstacles horizontally but also has been carefully adjusted vertically to ensure sufficient safety distance between the line and the ground and meet various engineering regulations.

[0111] This example combines DEM data with a local search optimization strategy to achieve three-dimensional elevation optimization of overhead transmission line paths. This approach not only ensures the feasibility of line construction on complex terrain but also reduces the need for conductor clearance above the ground by elevating nodes to local terrain high points, thereby improving the safety and stability of line operation. Furthermore, node-by-node elevation optimization ensures the economic rationality of the line under varying terrain conditions, providing a higher level of technical support for automated path planning.

[0112] Based on the above steps, determining the final path according to the updated plurality of fifth path nodes includes: sequentially determining horizontal distances between two adjacent fifth path nodes in the updated plurality of fifth path nodes, and determining whether any of the plurality of horizontal distances has a target horizontal distance greater than a third preset threshold; if it is determined that any target horizontal distance is greater than the third preset threshold, adding one or more intermediate path nodes between the two fifth path nodes corresponding to the target horizontal distances to obtain a plurality of seventh path nodes, such that the horizontal distances between any two adjacent seventh path nodes are less than or equal to the third preset threshold; sequentially determining slope values ​​of a pole-tower connection corresponding to two adjacent seventh path nodes in the plurality of seventh path nodes, and determining whether any of the plurality of slope values ​​has a target slope value greater than a fourth preset threshold; if it is determined that any target slope value is greater than the fourth preset threshold, adjusting the heights of the two seventh path nodes corresponding to the target slope values ​​to obtain a plurality of eighth path nodes, such that the slope values ​​corresponding to any two adjacent eighth path nodes are less than or equal to the fourth preset threshold; and determining the final path according to the plurality of eighth path nodes.

[0113] In this example, we explored how to further adjust the path based on the updated fifth path node (i.e., the node after preliminary elevation optimization) to meet engineering specifications for tower spacing and line slope, ultimately determining a feasible path for the overhead transmission line. This process, by adding intermediate nodes and adjusting node heights, ensures that the line is both safe and economical in actual construction.

[0114] 1. Horizontal distance inspection and addition of intermediate nodes:

[0115] Horizontal distance measurement: First, measure the horizontal distance between two adjacent fifth path nodes in the updated path. This step is to check whether the distance between towers is too long and does not meet the requirements of engineering specifications.

[0116] Third preset threshold: A third preset threshold is set (for example, the maximum allowable tower spacing is 300 meters) to determine whether the horizontal distance between adjacent nodes exceeds the threshold.

[0117] Adding intermediate nodes: If the horizontal distance to the target is greater than the third preset threshold, the algorithm will add one or more intermediate nodes (seventh path nodes) between the two nodes to ensure that the horizontal distance between any two adjacent nodes is less than or equal to the third preset threshold. The additional nodes also need to take terrain elevation and obstacles into consideration.

[0118] Node location selection: The location selection of intermediate nodes needs to be combined with DEM data to ensure that they are located in a legal construction area and are relatively flat to avoid unnecessary engineering difficulty and cost.

[0119] 2. Slope value check and node height adjustment:

[0120] Slope Calculation: Next, the algorithm calculates the slope of the line connecting the towers corresponding to two adjacent seventh-path nodes in the updated path. The slope reflects the line's inclination relative to the terrain and is a crucial indicator for ensuring safe and stable line operation.

[0121] Fourth preset threshold: A fourth preset threshold is set (for example, the maximum allowable slope is 5%) to verify whether the slope of the line is too large, affecting the project implementation and line performance.

[0122] Height Adjustment: If a target slope value is detected to be greater than the fourth preset threshold, the algorithm will fine-tune the height of the two seventh-path nodes involved in that slope value to adjust the slope to within the permitted range. Node height adjustment must be combined with DEM data to ensure that the adjusted node remains within the higher legal construction area.

[0123] Eighth Path Node Generation: After adjustment, the generated nodes are called eighth path nodes. This series of nodes forms a path that meets the dual constraints of minimum tower spacing and maximum allowable slope, further optimizing the engineering feasibility of the line.

[0124] 3. Final path determination:

[0125] Overall inspection and confirmation: The algorithm will eventually perform a comprehensive inspection of the entire path to ensure that the distances and slopes between all adjacent eighth path nodes meet the preset engineering specifications, including but not limited to tower spacing and line slope.

[0126] Path Determination: Once all conditions are met, the algorithm determines the final overhead transmission line path based on the location of the eighth path node. This path is not only rational in plan layout but also carefully tailored to vertical elevation and terrain slope, ensuring the safety and cost-effectiveness of the line during actual construction.

[0127] By adding nodes and fine-tuning heights based on horizontal distance limits and slope control, this embodiment provides a method for refining path planning and enhancing engineering adaptability. This method not only ensures the feasibility of line construction on complex terrain but also effectively controls line construction costs and operational risks through node adjustment and optimization, contributing to technological advancements in automated path planning and engineering design.

[0128] It should be noted that during the implementation of the above route planning method, the above route planning rules need to be followed, which specifically include but are not limited to:

[0129] 1. The path can cross linear features such as roads and rivers, but the path nodes (tower bases) must not be set on roads or rivers;

[0130] 2. Paths are strictly prohibited from crossing building areas, and path nodes cannot be located on buildings;

[0131] 3. When the path crosses a highway, the crossing angle should be as close to 90° (vertical crossing) as possible to reduce the length of the path along the highway and reduce safety hazards in line operation.

[0132] In an optional embodiment, the present application is combined with Figure 2 The implementation process of the above-mentioned dynamic planning method for intelligent line selection of transmission lines based on segmentation considerations is introduced. Figure 2 As shown, the specific steps include:

[0133] 2.1. Plan on the feature factor layer;

[0134] 2.2. Calling a Large Language Model to Generate Potential Point Routes: Based on the starting point, end point, and surrounding environment information, the API of a large language model (such as DeepSeek or LLM) is called, and interaction is performed using a few-shot prompting process. A list of potential pathpoints is output for each planning task. Environmental information can be provided as textual descriptions (e.g., topography, obstacle distribution) or structured data as model prompt input. The LLM then generates empirically recommended points, such as those near ridgelines and away from buildings (i.e., the potential pathpoint sequence described above).

[0135] 2.3. Perform sampling-based planning: Construct an exploration graph structure G for path growth (i.e., the exploration graph described above), add the starting point to graph G as the root node. Continue expanding the exploration graph G until the end point is found and the number of iterations does not exceed the maximum number of attempts N_max. The expansion process alternates between guided and random modes:

[0136] 2.3.1. Guided Expansion: Select the candidate point t with the highest priority from the potential point list as the target. Find the existing node p_near closest to t in the exploration graph G, and extend from p_near along the direction pointing to t for a certain number of steps to generate a new node p_new.

[0137] Collision Detection: Call the collisionFree function to check whether the line from p_near to p_new passes through any obstacles. If the line is collision-free within the current feature layer (i.e., no buildings blocking it and no inappropriate crossings along ground infrastructure), the extension is considered successful, p_new is added to graph G, and the connection between p_near and p_new is recorded. Otherwise, if a collision is detected (e.g., p_new falls within a building or the line passes through a building), the guided extension is aborted.

[0138] 2.3.2 Random Expansion: When guided expansion is blocked or random exploration is required according to a predefined strategy, a point p_rand is randomly sampled from the traversable free space. Similarly, the nearest node p_near in graph G is found and connected to obtain p_new. After collision detection to ensure there are no obstacles, p_new is added to graph G.

[0139] 2.3.3 Dynamic Adjustment: Whenever a new node is successfully added, the priority order of the remaining candidate points in the potential point list is updated in real time. For example, the priority of more promising points can be increased based on their proximity to the new node or whether they are already included in the current path, thereby improving the success rate of subsequent guided expansion.

[0140] Each time a new node is added, the distance from the node p_new to the endpoint is checked. If the distance is less than a preset threshold ϵ (indicating that the node is very close to the endpoint), the path is considered to be basically connected and the loop extension is terminated.

[0141] 2.4. Avoid and do not cross buildings, and try to cross roads perpendicularly: The following special constraints are followed during path planning:

[0142] 1) The path can cross linear features such as roads and rivers, but the path nodes (tower bases) must not be set on roads or rivers.

[0143] 2) Paths are strictly prohibited from passing through building areas, and path nodes cannot be located on buildings.

[0144] 3) When the path crosses a highway, the crossing angle should be as close to 90° (vertical crossing) as possible to reduce the length of the path along the highway and reduce safety hazards in line operation.

[0145] 2.5. Simplify the Road and Preserve Inflection Points: After the loop completes (either successfully reaching the endpoint or reaching the maximum number of iterations), a sequence of paths from the starting point to the endpoint is extracted from the exploration graph G. This path is then simplified by removing redundant intermediate nodes and retaining only key inflection points, ensuring that no straight lines between adjacent inflection points intersect with obstacles. Simplification is achieved by iteratively checking each segment of the path for collisions with obstacles, eliminating unnecessary bends, and ultimately obtaining an initial route with a streamlined tower layout.

[0146] 2.6 Planning on the DEM Data Layer: After completing the planar path planning for the feature layer and obtaining a preliminary path, the second stage of elevation optimization begins. This stage uses the digital elevation model (dem_merge data) to vertically adjust the locations of turning points in the path and, if necessary, add path points to accommodate terrain elevation differences.

[0147] 2.7. Fine-tune inflection points to higher altitudes: Obtain the altitude of each inflection point on the initial path (provided by the DEM). For each inflection point, search for local terrain high points in the vicinity and fine-tune the inflection point to a higher altitude location within the adjacent area that is legal for construction. This approach allows towers to be erected on hills or high ground as much as possible, increasing the line's clearance from the ground.

[0148] 2.8. Add and fine-tune the midpoints:

[0149] 2.8.1: Over-distance insertion of intermediate points: Measure the horizontal distance between adjacent towers. If the span exceeds the predetermined maximum threshold (i.e., the distance between the two towers is too long, which is not conducive to conductor suspension and mechanical stability), add an intermediate node between the two points. The selection principle of the newly added intermediate point is the same as above, preferably at a safe location with a higher altitude within the section, so as to simultaneously reduce the span and increase the support height;

[0150] 2.8.2 Slope Change Check: For adjacent pairs of tower nodes along the path, calculate the slope or vertical angle change of the line connecting them. If a section of the line is found to have a steep slope (for example, a line that drops sharply from a mountain to a valley), the position of the intermediate inflection point can be adjusted (slightly lowering or raising a point) to smooth the slope change and ensure that the line gradient meets engineering specifications.

[0151] After these adjustments, the final path is obtained. This final path is then verified to ensure that the terrain height in the middle of each section does not exceed that at the segment endpoints. In other words, the ground elevation at the center of any span does not exceed the elevation of the tower foundations at either end. This approach optimizes the entire line in three dimensions, avoiding obstacles while adapting to the terrain, resulting in an optimized path that meets safety distance and installation requirements.

[0152] Through the description of the above embodiments, those skilled in the art can clearly understand that the methods according to the above embodiments can be implemented by means of software plus the necessary general hardware platform. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application, or the part that contributes to the existing technology, can be embodied in the form of a software product. The computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0153] The present application also provides a multi-stage route planning device based on large model recommendation and terrain optimization, such as Figure 3 As shown, Figure 3 This is a structural block diagram of an optional multi-stage route planning device based on large model recommendation and terrain optimization according to an embodiment of the present application, the device comprising:

[0154] a generation module 32 for generating a potential pathpoint sequence based on the start point information, the end point information, and environmental data using a large language model, wherein the start point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located;

[0155] A search module 34 is configured to perform a two-dimensional path search on the potential path point sequence on the feature layer using a target algorithm to generate an initial path of the overhead transmission line;

[0156] The optimization module 36 is configured to perform three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line.

[0157] Through the above-mentioned device, the starting point information, the end point information and the environmental data are first processed by the large language model to generate a sequence of potential path points, wherein the starting point information and the end point information are respectively used to indicate the starting point and the end point of the overhead transmission line to be constructed, and the environmental data are used to indicate the environmental information in the target area where the overhead transmission line is located; then the target algorithm is used to recommend potential path points and quickly grow the path to obtain the initial path; finally, the generated initial path is fine-tuned in elevation and supplemented with nodes, thereby completing the fine optimization of the line and obtaining the final path; adopting the above-mentioned scheme, the large language model and the environmental understanding ability are combined with the path search algorithm, and the digital elevation model is introduced in stages for optimization to realize fully automatic line selection and planning of overhead lines; through this technical scheme, the line selection efficiency can be significantly improved, the dependence on manual experience can be reduced, and a safer, more reliable and economically reasonable line plan can be generated; thereby solving the problem in the related technology that the line selection and planning process of overhead transmission lines is affected by many factors and they restrict each other, manual line selection is time-consuming and labor-intensive, the planning efficiency is low and the line plan quality is poor.

[0158] In an exemplary embodiment, the above-mentioned search module 34 is also used to initialize and generate an exploration graph, wherein the starting point information is the root node of the exploration graph; multiple candidate points in the potential path point sequence whose distance from the root node is less than a first preset threshold are determined as an initial priority sequence, wherein the multiple candidate points are used to indicate multiple path directions of the initial path; the exploration graph is cyclically expanded through a potential point priority mechanism and / or a random node sampling method until the node in the exploration graph is connected to the path node corresponding to the end point information or the number of cyclic extensions reaches a preset number of iterations; a path sequence is extracted from the exploration graph according to the root node and the path node corresponding to the end point information, and the initial path is generated according to the path sequence.

[0159] In an exemplary embodiment, the above-mentioned search module 34 is also used to determine the second path node with the highest priority among the first path nodes when it is determined that there is a first path node in the potential path point sequence that has not been added to the exploration graph; determine whether the subpath between the third path node newly added in the exploration graph and the second path node complies with the route planning rules, wherein the route planning rules are used to indicate that the initial path does not overlap with the area where the obstacle is located, and the subpath is a straight line path; when it is determined that the route planning rules are complied with, the second path node is added to the exploration graph; when it is determined that the first path node does not exist or is not complied with, a fourth path node is randomly selected from the first sub-area from the third path node to the path node corresponding to the end point information by the random node sampling method, and when it is determined that the subpath between the third path node and the fourth path node complies with the route planning rules, the fourth path node is added to the exploration graph, wherein the target area includes the first sub-area.

[0160] Furthermore, the search module 34 is further configured to increase the sampling frequency of the random node sampling method when the number of failures in expanding the exploration graph through the potential point priority mechanism reaches a second preset threshold.

[0161] Optionally, the above-mentioned optimization module 36 is also used to identify non-critical nodes of multiple path nodes in the initial path to obtain multiple non-critical nodes, wherein the sub-path where the non-critical nodes are located does not overlap with the area where the obstacle is located, and the sub-path is a straight path; the multiple non-critical nodes are deleted from the initial path to obtain a simplified path; and the path nodes in the simplified path are optimized in three-dimensional elevation according to the digital elevation data to obtain the final path.

[0162] In an exemplary embodiment, the above-mentioned optimization module 36 is also used to determine the altitudes of multiple fifth path nodes in the simplified path based on the digital elevation data; for the sixth path node among the multiple fifth path nodes, search for a local terrain high point in the second sub-area where the sixth path node is located, and update the sixth path node to the local terrain high point to obtain multiple updated fifth path nodes, wherein the local terrain high point is a legal construction area; and determine the final path based on the updated multiple fifth path nodes.

[0163] Optionally, the optimization module 36 is further configured to sequentially determine horizontal distances between two adjacent fifth path nodes among the updated plurality of fifth path nodes, and determine whether any of the plurality of horizontal distances has a target horizontal distance greater than a third preset threshold; if it is determined that any target horizontal distance is greater than the third preset threshold, add one or more intermediate path nodes between the two fifth path nodes corresponding to the target horizontal distances to obtain a plurality of seventh path nodes, such that the horizontal distances between any two adjacent seventh path nodes are less than or equal to the third preset threshold; sequentially determine slope values ​​of the pole-tower connection corresponding to two adjacent seventh path nodes among the plurality of seventh path nodes, and determine whether any of the plurality of slope values ​​has a target slope value greater than a fourth preset threshold; if it is determined that any target slope value is greater than the fourth preset threshold, adjust the heights of the two seventh path nodes corresponding to the target slope value to obtain a plurality of eighth path nodes, such that the slope values ​​corresponding to any two adjacent eighth path nodes are less than or equal to the fourth preset threshold; and determine the final path based on the plurality of eighth path nodes.

[0164] An embodiment of the present application further provides a storage medium, which includes a stored program, wherein the program executes any of the above methods when it is run.

[0165] Optionally, in this embodiment, the storage medium may be configured to store program codes for executing the following steps:

[0166] S1, generating a potential pathpoint sequence based on starting point information, end point information, and environmental data using a large language model, wherein the starting point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located;

[0167] S2, performing a two-dimensional path search on the potential path point sequence using a target algorithm on the feature layer to generate an initial path of the overhead transmission line;

[0168] S3, performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line.

[0169] An embodiment of the present application further provides an electronic device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0170] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0171] Optionally, in this embodiment, the processor may be configured to execute the following steps through a computer program:

[0172] S1, generating a potential pathpoint sequence based on starting point information, end point information, and environmental data using a large language model, wherein the starting point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located;

[0173] S2, performing a two-dimensional path search on the potential path point sequence using a target algorithm on the feature layer to generate an initial path of the overhead transmission line;

[0174] S3, performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line.

[0175] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0176] An embodiment of the present application further provides a computer program product, comprising a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program product, and when the computer program is executed by a processor, the steps of the method described in each embodiment of the present application are implemented.

[0177] Optionally, in this embodiment, the computer program may be configured to implement the following steps when executed by a processor:

[0178] S1, generating a potential pathpoint sequence based on starting point information, end point information, and environmental data using a large language model, wherein the starting point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located;

[0179] S2, performing a two-dimensional path search on the potential path point sequence using a target algorithm on the feature layer to generate an initial path of the overhead transmission line;

[0180] S3, performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line.

[0181] Optionally, specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0182] Obviously, those skilled in the art should understand that the modules or steps of the present application described above can be implemented using a general-purpose computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices. Alternatively, they can be implemented using program code executable by the computing device, so that they can be stored in a storage device and executed by the computing device. In some cases, the steps shown or described can be performed in a different order than herein, or they can be made into separate integrated circuit modules, or multiple modules or steps can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0183] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application shall be included within the scope of protection of the present application.

Claims

1. A dynamic planning method for intelligent line selection of power transmission lines based on segmentation considerations, characterized in that: include: Generate a potential pathpoint sequence based on the starting point information, the ending point information, and environmental data using a large language model, wherein the starting point information and the ending point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located; Performing a two-dimensional path search on the potential path point sequence on the feature layer using a target algorithm to generate an initial path for the overhead transmission line; Performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line; The method of performing a two-dimensional path search on the potential path point sequence by using a target algorithm on the feature layer to generate an initial path of the overhead transmission line includes: Initialize and generate an exploration graph, wherein the starting point information is the root node of the exploration graph; Determining a plurality of candidate points in the potential path point sequence whose distances from the root node are less than a first preset threshold as an initial priority sequence, wherein the plurality of candidate points are used to indicate a plurality of path directions of the initial path; Cyclic expansion of the exploration graph is performed using a potential point priority mechanism and / or a random node sampling method until a node in the exploration graph is connected to a path node corresponding to the endpoint information or the number of cyclic expansions reaches a preset number of iterations; Extracting a path sequence from the exploration graph according to the path nodes corresponding to the root node and the end point information, and generating the initial path according to the path sequence; The exploration graph is cyclically expanded by a potential point priority mechanism and / or a random node sampling method, including: In a case where it is determined that there is a first path node in the potential path point sequence that has not been added to the exploration graph, determining a second path node with the highest priority among the first path nodes; Determining whether a subpath between a third path node newly added to the exploration graph and the second path node complies with a route planning rule, wherein the route planning rule is used to indicate that the initial path does not overlap with an area where an obstacle is located, and the subpath is a straight path; If it is determined that the route planning rule is met, adding the second path node to the exploration graph; When it is determined that the first path node does not exist or is determined not to comply with the route planning rules, a fourth path node is randomly selected from the first sub-area from the third path node to the path node corresponding to the end point information by the random node sampling method, and when it is determined that the subpath between the third path node and the fourth path node complies with the route planning rules, the fourth path node is added to the exploration graph, wherein the target area includes the first sub-area.

2. The method for intelligent transmission line selection dynamic planning based on segmentation consideration according to claim 1 is characterized in that: The method further comprises: When the number of failures in expanding the exploration graph by the potential point priority mechanism reaches a second preset threshold, the sampling frequency of the random node sampling method is increased.

3. The method for intelligent transmission line selection dynamic planning based on segmentation consideration according to claim 1 is characterized in that: Before performing three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain the final path of the overhead transmission line, the method further includes: Identifying non-critical nodes on multiple path nodes in the initial path to obtain multiple non-critical nodes, wherein a subpath where the non-critical nodes are located does not overlap with an area where an obstacle is located, and the subpath is a straight path; Deleting the plurality of non-critical nodes from the initial path to obtain a simplified path; The path nodes in the simplified path are optimized in three dimensions according to the digital elevation data to obtain the final path.

4. The method for intelligent transmission line selection dynamic planning based on segmentation consideration according to claim 3 is characterized in that: Performing three-dimensional elevation optimization on the path nodes in the simplified path according to the digital elevation data to obtain the final path includes: determining the altitudes of a plurality of fifth path nodes in the simplified path based on the digital elevation data; For a sixth path node among the plurality of fifth path nodes, searching for a local terrain high point in the second sub-area where the sixth path node is located, and updating the sixth path node to the local terrain high point to obtain a plurality of updated fifth path nodes, wherein the local terrain high point is a legal construction area; The final path is determined according to the updated plurality of fifth path nodes.

5. The method for dynamic planning of intelligent transmission line selection based on segmentation consideration according to claim 4 is characterized in that: Determining the final path according to the updated plurality of fifth path nodes includes: sequentially determining horizontal distances between two adjacent fifth path nodes among the updated plurality of fifth path nodes, and determining whether a target horizontal distance among the plurality of horizontal distances is greater than a third preset threshold; If it is determined that the target horizontal distance is greater than the third preset threshold, one or more intermediate path nodes are added between two fifth path nodes corresponding to the target horizontal distance to obtain multiple seventh path nodes, so that the horizontal distance between any two adjacent seventh path nodes is less than or equal to the third preset threshold; Determining in sequence the slope values ​​of the pole-tower connection corresponding to two adjacent seventh path nodes among the plurality of seventh path nodes, and determining whether there is a target slope value greater than a fourth preset threshold value among the plurality of slope values; If it is determined that the target slope value is greater than the fourth preset threshold, adjusting the heights of the two seventh path nodes corresponding to the target slope value to obtain a plurality of eighth path nodes, such that the slope values ​​corresponding to any two adjacent eighth path nodes are less than or equal to the fourth preset threshold; The final path is determined according to the plurality of eighth path nodes.

6. A multi-stage route planning device based on large model recommendation and terrain optimization, characterized in that: include: a generation module, configured to generate a potential pathpoint sequence based on start point information, end point information, and environmental data using a large language model, wherein the start point information and the end point information are used to construct an overhead transmission line, and the environmental data is used to indicate environmental information of a target area where the overhead transmission line is located; A search module is used to perform a two-dimensional path search on the potential path point sequence on the feature layer by using a target algorithm to generate an initial path of the overhead transmission line; an optimization module, configured to perform three-dimensional elevation optimization on the path nodes in the initial path according to the digital elevation data to obtain a final path of the overhead transmission line; Wherein, the search module is also used to initialize and generate an exploration graph, wherein the starting point information is the root node of the exploration graph; multiple candidate points in the potential path point sequence whose distance from the root node is less than a first preset threshold are determined as an initial priority sequence, wherein the multiple candidate points are used to indicate multiple path directions of the initial path; the exploration graph is cyclically expanded through a potential point priority mechanism and / or a random node sampling method until a node in the exploration graph is connected to a path node corresponding to the end point information or the number of cyclic expansions reaches a preset number of iterations; a path sequence is extracted from the exploration graph according to the root node and the path node corresponding to the end point information, and the initial path is generated according to the path sequence; Wherein, the search module is also used to determine the second path node with the highest priority among the first path nodes when it is determined that there is a first path node in the potential path point sequence that has not been added to the exploration graph; determine whether the subpath between the third path node newly added in the exploration graph and the second path node complies with the route planning rules, wherein the route planning rules are used to indicate that the initial path does not overlap with the area where the obstacle is located, and the subpath is a straight line path; when it is determined that the route planning rules are complied with, the second path node is added to the exploration graph; when it is determined that the first path node does not exist or is not complied with, a fourth path node is randomly selected from the first sub-area from the third path node to the path node corresponding to the end point information by the random node sampling method, and when it is determined that the subpath between the third path node and the fourth path node complies with the route planning rules, the fourth path node is added to the exploration graph, wherein the target area includes the first sub-area.

7. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the method according to any one of claims 1 to 5 is executed when the program is executed.

8. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images