A method for planning a temporary road for a power transmission line in view of multi-tower cooperative planning

Abstract

The application discloses a kind of temporary road planning methods for multi-tower collaborative planning of transmission line, and it relates to the technical field of power engineering construction.The method comprises: obtaining transmission corridor features, sensitive area, tower site coordinates and DEM data and constructing a grid map;Based on cluster analysis, the virtual end point of each group is extracted from the existing road;Through grid corridor analysis and multi-dimensional vector aggregation, the virtual starting point of each group is generated;Improved A algorithm is used to plan the main path from the virtual starting point to the existing road and update the grid map;According to the accessibility of tower site to virtual starting point, the dynamic growing built road endpoint set is sequentially targeted, and the improved A algorithm is used to plan the branch path of each tower site;The built road endpoint set is inherited between each group.The application realizes the iterative reuse of multi-tower collaborative planning and built road, reduces the length of temporary road construction, reduces the construction cost, and improves the path planning quality and engineering feasibility.

Description

Technical Field

[0001] This invention relates to the field of power engineering construction technology, specifically to a method for planning temporary roads for transmission lines oriented towards multi-tower collaborative planning. Background Technology

[0002] In the mechanized construction of power transmission lines, temporary construction roads are a necessary condition for the entry of large construction machinery (such as heavy cranes, rotary drilling rigs, tensioners, etc.). The planning quality of temporary roads not only directly affects whether construction machinery can safely reach each tower location, but their construction costs (including earthwork, compensation for crops, and occupation of forest land) are usually a key aspect of project cost control.

[0003] As power grid projects extend into mountainous and hilly terrain, a single transmission line often involves dozens or even hundreds of tower sites, making the need for collaborative planning across multiple tower sites increasingly prominent. However, existing temporary road planning methods still have the following shortcomings when dealing with scenarios involving multiple tower sites and complex terrain: First, the planning model lacks a multi-tower collaborative mechanism, resulting in low road reuse rates. Existing technologies typically employ a "single-tower independent planning" or "nearest access" strategy, meaning that each tower independently solves for its shortest or nearest path to the existing road network, with the planning processes for each tower isolated and unrelated. This strategy leads to a "radial" or "comb-like" distribution pattern in the temporary road network, with road segments that could be shared between adjacent towers being repeatedly planned, resulting in a large number of parallel or even intersecting redundant road segments. This results in low road reuse rates and a significant increase in earthwork and crop compensation costs.

[0004] Second, the ability to quantify constraints related to complex terrain is insufficient, resulting in poor on-site feasibility of the planning results. Mechanized construction places strict technical requirements on indicators such as the slope, turning radius, and foundation bearing capacity of temporary roads. Existing technologies largely rely on manual on-site surveys combined with two-dimensional GIS mapping, making it difficult to systematically integrate multi-dimensional constraints such as elevation gradient, terrain curvature, and distribution of economic crops during the planning stage. Especially in complex mountainous areas, manually planned routes often suffer from problems such as excessive local slopes and too many sharp bends, leading to frequent modifications to the alignment during construction, increasing rework costs and safety risks.

[0005] Third, the computational complexity of multi-tower joint planning is high, making it difficult to balance efficiency and quality. From a computational theory perspective, multi-tower joint road planning is essentially a constrained Steiner's minimum tree problem (NP-hard problem). Traditional exhaustive search methods experience exponential growth in computation time as the number of towers increases, rendering them impractical for engineering applications. While conventional heuristic algorithms can shorten computation time, they are prone to getting trapped in local optima and struggle to simultaneously address multiple optimization objectives such as shortest path, optimal slope, and avoidance of sensitive areas, making it difficult to achieve a balance between computational efficiency and planning quality.

[0006] In summary, existing technologies urgently need a temporary road planning method with a multi-tower collaborative mechanism and dynamic road network growth capability to overcome the limitations of independent planning by a single tower and achieve global optimization and cost minimization of the temporary road network. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a method for planning temporary roads for transmission lines in a multi-tower collaborative planning manner, which can reduce the length of temporary roads and reduce the cost of road construction.

[0008] The technical solution adopted by this invention to solve its technical problem is a method for planning temporary roads for mechanized construction of transmission lines oriented towards multi-tower collaborative planning, comprising the following steps: S1. Acquire data on terrain features, sensitive areas, tower coordinates, and DEM elevation within the power transmission corridor, and set a minimum permissible turning radius threshold. Maximum allowable slope and road width And construct a grid map and determine the tower clusters. At the same time, construct a set of grid points for existing roads. ; S2. To Each group ,by The predetermined range is extended to the front, back, left, and right sides of the inner tower location. The inner tower raster map is extracted from the raster map, and the virtual endpoint in the inner tower raster map is obtained using a clustering analysis-based method. ; S3. To Each group The virtual starting point is obtained by using a grid corridor analysis and multi-dimensional vector aggregation method. ; S4. Adopting improved A Algorithm obtained to Optimal path grid set And update the raster map, The middle grid update is the endpoint, with All grid points and All elements in the current group Collection of the ends of existing roads The initial element; S5. Calculation Each tower to Find the reachable vector magnitudes and sort them from low to high to obtain the sorted vectors. ; S6. From At the beginning, for Repeat the following steps for all elements: Starting from the current set of existing road endpoints on the raster map. As the endpoint, adopt improved A The algorithm obtains the paths to all tower locations. And update the raster map, Updated to endpoint and added ; S7. Traversal Each group within the group repeats S2-S6, and upon entering the next group, the previous group's... As the initial element, it is inherited into the current group's endpoint set until completion. Search all tower location paths.

[0009] Furthermore, S2 includes the following steps: S201. Three-dimensional feature vector construction: Extracting the existing road set from the inner tower grid map. Intersecting grids serve as inner tower road grids. Each inner tower road grid is represented as a point in three-dimensional space, and its feature vector is... ; S202. Perform Z-score standardization so that the mean of each dimension is 0 and the standard deviation is 1; S203. Perform 3D DBSCAN clustering to obtain cluster labels and intra-cluster road points. ; S204. Generate a virtual endpoint: (This is related to...) For each cluster, calculate the feature vector of all road points within it. The arithmetic mean of each dimension is used as the three-dimensional centroid coordinates of the cluster, serving as the virtual endpoint. .

[0010] Furthermore, S3 includes the following steps: S301. To Multidimensional feature extraction was performed on each tower location: Constructing tower sites To the finish line Straight corridor area This region contains all lines. left and right sides The distance of each grid cell, where The value is a straight line Length ; Calculate horizontal distance ; Calculate obstacle density ; Calculate the density of cash crops ; Calculate the mean absolute elevation difference ; Calculate the maximum elevation difference ; Calculate the rate of change of elevation ; Calculate the degree of change in elevation gradient ,in, refer to The number of grid cells, Refers to obstacle grids, Refers to the grid of economic crops. This refers to the elevation difference between adjacent grid cells. This is an indicator function; it takes a value of 1 when two adjacent elevation changes are in opposite directions, and 0 otherwise. and These represent the maximum and minimum elevation values ​​within the region, respectively. These are the weighting coefficients; like ,in If a preset obstacle threshold is not met, the corridor is determined to be impassable, and the vector is invalid. S302. Construction of Multi-Dimensional Reachable Vectors: [The sentence is incomplete and likely refers to a separate topic Other tower locations within the cluster Construct reachable vectors for adjacent tower locations to the endpoint. , with virtual endpoint set Construct a destination reachable vector from each point in the vector. Based on the corridor analysis method of S301, the traversability of each vector is determined one by one, and a set is formed from all traversable vectors. ; S303. Calculate multidimensional modulus For any vector The formula for calculating its modulus is: , in The length of the diagonal of the straight corridor area. , , , ; S304. Reachability Vector Filtering and Aggregation: For each starting point Calculate the minimum effective modulus Retain all that meet the requirements The reachable vectors form the filtered set. Merging the filtering vectors from all starting points yields the global vector set. Extract the start and end coordinates of all vectors to form a node set. ; S305. Closed-Loop Polygon Detection and Topology Analysis: For Node Sets Construct an adjacency matrix, use depth-first search to detect closed polygons, and filter out polygons with areas smaller than a preset threshold, or those with self-intersections or duplicate vertices to obtain an effective set. ; S306. Virtual Starting Point Estimation: like Calculate the geometric centroid of an element with a quantity of 1. ,in The number of vertices; like If the number of elements is greater than 1, calculate the geometric centroid of each polygon, merge the set of centroids with the virtual endpoint, and use the Graham scan algorithm to calculate the convex polygon and find its centroid. ; At the center of mass around Within the neighborhood, calculate each passable grid cell. Overall rating , in For elevation, The average elevation of the neighborhood. For elevation standard deviation, Indicator values ​​for cash crops, The distance from the grid to the centroid. These are preset coefficients; The grid with the best overall score is used as the virtual starting point. .

[0011] Furthermore, the improvement A in S4 The algorithm includes a constraint-aware node state expansion and verification mechanism, and the steps are as follows: S401. An 8-neighborhood expansion strategy is used to generate candidate neighbor nodes for each current node, and three-level constraint verification is implemented: S401-1. Level 1 Obstacle Avoidance Verification: Check if the candidate node attribute is an obstacle; if so, remove it. S401-2. Second-level slope verification: Calculate the current node. to candidate node slope value ,in The elevation difference between adjacent grid cells. The grid width (i.e., road width) is... Then remove; S401-3. Level 3 Turning Radius Verification: Based on Parent Node Current node and candidate nodes The three-node sequence is used to calculate the turning radius. : Calculate the horizontal projection vector , , Calculate the angle of change of direction , in , , ; Calculate the turning radius ,in ,like Then remove it.

[0012] Furthermore, the improvement A in S4 The algorithm also includes the following steps: S402. Multi-terminal heuristic function, the specific steps are as follows: S402-1. Topological reachability pre-screening of virtual endpoints: Obtain the connected component labels of the current node, remove virtual endpoints with inconsistent connected component labels, and perform a secondary screening based on the reachability calculation method in S302 to obtain the set of valid endpoints. ; S402-2. Dynamic Heuristic Function Calculation: Traversal The elements in the array, for each endpoint Calculate the basic moving cost Nonlinear terrain correction ,in This is the terrain correction factor. The actual slope from the starting point to the end point. The optimal slope; Economic penalties , in For the grid cost coefficient of cash crops, This represents the number of economic crop grid cells within a straight corridor. This represents the total number of grid cells in the corridor. Detour penalty ,in Number of obstacle grids; composite heuristic value ; S402-3. Heuristic Value Composition and Guidance for Multiple Endpoints: Take the minimum composite heuristic value among all valid endpoints as the heuristic value of the current node. .

[0013] Furthermore, the improvement in S6 is A The algorithm uses the sorted tower position sequence The order is based on the current set of completed road endpoints. For multi-destination path search, after completing the path planning for each tower location, all grid nodes of that path are added. This allows the endpoint set to grow dynamically.

[0014] The beneficial effects of this invention are: 1. This invention constructs a two-stage hierarchical decoupled architecture from main line access to branch line convergence by combining the steps of virtual starting point generation, main line path planning, and branch line path planning. This mechanism breaks through the traditional flattened strategy of independent planning or nearby access for single towers, enabling the temporary road network to naturally form a tree-like topology structure with virtual hubs as the core. It avoids the radial or comb-like redundant road sections caused by multiple towers independently accessing the existing road network, and realizes the sharing of temporary access roads by multiple towers, thereby reducing the road construction length and corresponding earthwork, crop compensation, and other costs from the source.

[0015] 2. In the branch line path planning process of this invention, after the path planning for each tower location is completed, all grid nodes of that path are added to the set of existing road endpoints, allowing the endpoint set to grow dynamically as the planning progresses. Subsequent tower locations use this dynamic set as multiple endpoints for path searching, automatically identifying and connecting to previously constructed branch lines. Simultaneously, through an endpoint set inheritance mechanism between clusters, existing roads across clusters can also be reused. This mechanism endows the road network with dynamic growth capabilities, overcoming the deficiency in traditional static planning where subsequent paths cannot utilize previous results, promoting spatial merging and sharing between branch lines, and significantly reducing the workload caused by redundant planning. Attached Figure Description

[0016] Figure 1 This is a flowchart of the present invention; Figure 2 This is a schematic diagram of the tower cluster; Figure 3 This is the virtual starting point calculation process for group 5 in this embodiment of the invention; Figure 4 This is the virtual starting point calculation process for group 12 in this embodiment of the invention; Figure 5 This is the virtual starting point calculation process for group 14 in this embodiment of the invention; Figure 6 This is the path planning process for each tower location in Group 5 according to an embodiment of the present invention; Figure 7 This is the path planning result for each tower location in Group 5 of this invention. Detailed Implementation

[0017] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0018] The transmission line project involved in this embodiment includes 32 tower sites. The terrain along the route is mainly hilly, and the existing road network consists of county-level highways and rural roads, which are sparsely distributed and of low grade. The construction requires the use of mechanized equipment, and the technical requirements for the temporary roads are as follows: maximum slope ≤15°, minimum turning radius ≥6.5m, and road width 3.5m.

[0019] See Figure 1 The specific embodiments of the present invention will be described in detail below.

[0020] S1. Data Acquisition and Preprocessing

[0021] S1-1. Basic Data Collection and Storage

[0022] The multi-source data within the power transmission corridor includes: (1) land feature type data, including land feature classification layers such as buildings, water systems, roads, forest land, cultivated land (including economic crops), and grassland, all of which are vector area data; (2) sensitive area data, including spatial range data such as nature reserves, water source protection areas, basic farmland, and ecological red lines; (3) tower location coordinate data, which are the center pile coordinates of each tower location determined during the design phase of this line project, including plane coordinates and design pile top elevation, stored in Shapefile point file format; (4) DEM elevation data, using a 1:2000 scale digital elevation model, with a grid cell size of 2m×2m, and the elevation accuracy meets the secondary accuracy requirements in the "Quality Inspection Standard for Digital Elevation Models" (CH / T 1029-2012).

[0023] S1-2. Road Parameter Settings

[0024] The construction requires the use of mechanized equipment, and the technical requirements for the temporary road are as follows: maximum slope ≤ 15°, minimum turning radius ≥ 6.5m, and road width 3.5m.

[0025] S1-3. Raster Map Construction

[0026] All the above vector data were converted to raster format. The basic cell size of the raster map is [size missing]. m. The specific conversion method is as follows: (1) Determine the spatial extent of the raster map: Based on the bounding rectangle of all tower coordinates, extend 1km outwards as the map boundary to ensure coverage of all possible route planning areas; (2) Rasterization of each layer: The "majority rule" resampling method is used to convert vector surface data such as land cover type and sensitive area into raster layers, and each raster cell stores a category code value; (3) DEM resampling: The original 0.5m resolution DEM was resampled to 3.5m resolution using bilinear interpolation, and the elevation value of each raster cell was stored. ; (4) Multi-layer raster overlay: Overlay the land feature type raster, sensitive area raster, and DEM elevation raster under the same spatial reference to form a comprehensive raster map containing multi-attribute information.

[0027] S1-4. Tower Site Grouping

[0028] See Figure 2 The designers divide the tower sites into clusters based on the site survey (considering existing roads, land features, sensitive areas, and other site conditions), thus forming a collection of tower site clusters. In this embodiment, the 32 tower sites are divided into 16 clusters, and each cluster contains 1 to 4 spatially adjacent tower sites. The grouping results are recorded using a correspondence table between cluster numbers and tower site numbers.

[0029] S1-5. Construction of Existing Road Grid Point Set

[0030] Extract all road categories, including national highways, provincial highways, county roads, township roads, and farm roads, from the land cover type data. After rasterizing this vector road data, extract the row and column coordinates of all road raster cells to form a set of existing road raster points. ,in This represents the total number of road grid points. Each road grid point... Includes planar coordinates and elevation value .

[0031] S2. Virtual endpoint generation

[0032] Grouping each tower location The most representative access points are extracted from the existing roads surrounding the cluster as virtual endpoints. This provides multiple candidate targets for subsequent main path search. (Based on grouping) Let's take an example to explain in detail.

[0033] by Using the lines connecting all the tower locations as a baseline, a rectangular local area is formed by extending 500m in each of the forward and backward directions of the line, as well as to the left and right sides. The raster data of this local area is cropped from the global raster map and denoted as the inner tower raster map. The inner tower grid map contains all grid information, including tower location, existing roads, terrain, and features.

[0034] S201. Construction of 3D Feature Vectors

[0035] From the inner tower grid diagram Extract from existing road sets Intersecting grid cells are used as inner tower road grid cells. The specific determination method is: traversal... All grids in the set, if the spatial coordinates of a certain grid are... and Spatial coordinates of any road grid satisfy If so, then the grid is marked as an inner tower road grid.

[0036] For all inner tower road grids, construct a three-dimensional feature vector. ,in This represents the horizontal coordinate of the grid (in meters). The vertical coordinate is the plane coordinate (unit: m). This represents the elevation value (in meters) of the grid location, with all three dimensions expressed as floating-point numbers.

[0037] S202. Z-score standardization

[0038] All 3D feature vectors constructed in S201 are Z-score normalized using the following formula:

[0039] in For all inner tower road grids Mean across three dimensions These represent the corresponding standard deviations. After standardization, the mean of each dimension is 0 and the standard deviation is 1, eliminating the influence of dimensional differences on the clustering results.

[0040] S203. 3D DBSCAN Clustering

[0041] The DBSCAN clustering algorithm is performed on the standardized 3D feature points. In this embodiment, the clustering parameter is set as: neighborhood radius. m, minimum number of points .

[0042] The execution steps of the DBSCAN algorithm are as follows: (1) For each unvisited point, search its - All points within the neighborhood; (2) If the number of points within the neighborhood If a new cluster is created, all points in the neighborhood are added to the cluster; (3) the cluster is recursively expanded, and all points with reachable density are added to the same cluster; (4) points that cannot be assigned to any cluster are marked as noise points.

[0043] In this embodiment, The inner tower road grid was clustered into 3 clusters, denoted as Each cluster corresponds to a relatively concentrated access section of the existing road network around the cluster. Noise points are discarded and do not participate in subsequent virtual endpoint calculations.

[0044] S204. Virtual endpoint generation

[0045] For each cluster Calculate its three-dimensional centroid coordinates as the virtual endpoint. :

[0046] in For clusters The number of points along the middle road.

[0047] The centroid of all clusters constitutes virtual endpoint set These three virtual endpoints represent the group. There are existing road access opportunities in three different directions around the area.

[0048] S3. Virtual Starting Point Generation

[0049] The purpose of this step is to group each tower location together. Through grid corridor analysis and multi-dimensional vector aggregation, an optimal "convergence hub" location is automatically identified as the virtual starting point. This virtual starting point will serve as the convergence point for all branch paths within the cluster, as well as the starting point for the main path, achieving a two-stage hierarchical decoupling of "branch convergence → main path access".

[0050] This embodiment takes Group 5 (including towers T6, T7, and T8, located in the transition zone from hills to plains), Group 12 (including towers T20, T21, T22, and T23), and Group 14 (including towers T26, T27, and T28, located in the ridge area) as examples to illustrate the virtual starting point calculation process under different terrain conditions.

[0051] S301. Multidimensional Feature Extraction

[0052] S301-1. Construction of Straight Corridor Areas: For Clusters Any tower position in To any destination Construct a straight corridor area The endpoint It can be other tower locations within the group, or it can be a virtual endpoint collection. A point in the middle.

[0053] Straight corridor area The specific construction method is as follows: (1) Calculate from arrive The direction vector of the line (2) Calculate the corridor width parameters That is, the width of the corridor adapts to the spacing between towers, and the wider the spacing, the wider the corridor, so as to cover more terrain information; (3) Traverse all grids in the grid map. Calculate the grid to the line vertical distance ,like Then the grid will be included in the corridor area. (4) The number of grid cells is denoted as .

[0054] S301-2. Calculation of features in various dimensions within the corridor: (1) Horizontal distance Tower position and the finish line The Euclidean distance between them is calculated using the following formula: The unit is meters (m). This value reflects the basic length cost of the path.

[0055] (2) Obstacle density The proportion of obstacle grids within the corridor area, including inaccessible features such as buildings, water bodies, steep cliffs, military facilities, and the core area of ​​nature reserves, is calculated using the following formula: ,in For a binary indicator function, if the grid If it is an obstacle, the value is 1; otherwise, it is 0. The range of values ​​is The larger the value, the more difficult it is to pass through the corridor.

[0056] (3) Density of cash crops The proportion of economic crop grids within the corridor area, including orchards, tea gardens, economic forests, medicinal herb cultivation areas, greenhouses, etc., is calculated using the following formula: ,in For a binary indicator function, if the grid If it is a cash crop, the value is 1; otherwise, it is 0. The range of values ​​is The larger the value, the greater the impact of the path on the agricultural economy, and the higher the corresponding cost.

[0057] (4) Elevation gradient characteristics: ① Mean absolute elevation difference The average absolute value of the elevation difference between adjacent grid cells within the corridor reflects the overall topographic relief of the corridor. The calculation formula is as follows: ,in The elevation difference between adjacent grid cells along the corridor direction (unit: m).

[0058] ② Maximum elevation difference The maximum absolute value of the elevation difference between adjacent grid cells within the corridor is used to identify extreme terrain such as steep slopes and cliffs. The calculation formula is as follows: .

[0059] ③ Rate of change of elevation The frequency with which two adjacent elevation changes within a corridor are in opposite directions (i.e., the terrain changes from rising to falling or from falling to rising) reflects the degree of terrain curvature. A higher frequency indicates more complex and undulating terrain. The calculation formula is as follows: ,in For indicator functions, when The value is 1 when the elevation changes in the opposite direction, and 0 otherwise.

[0060] ④ Degree of change in elevation gradient The weighted sum of the three elevation features mentioned above is calculated using the following formula:

[0061] in and These represent the maximum and minimum elevation values ​​for all grid cells within the corridor area, respectively. The elevation range of the corridor area is used to normalize the first two terms. The weighting coefficient in this embodiment is set to [value missing]. , , These correspond to the relative importance of the three factors: overall undulation, local steep slope, and topographical curvature.

[0062] S301-3. Pipeline Verification: If In this embodiment, In other words, if obstacles account for more than 40% of a corridor, it is considered impassable, and the corresponding vector is marked as invalid and not included in subsequent calculations. This threshold strikes a balance between path feasibility and detour cost—a threshold that is too small will lead to many corridors being falsely judged as impassable, while a threshold that is too large may include severely obstructed paths.

[0063] S302. Construction of Multi-Dimensional Reachable Vectors

[0064] Group Each tower position Construct two types of reachable vectors respectively: (1) Reachable vector of adjacent tower sites : other tower positions within the group ( Using ) as the endpoint, construct from arrive straight corridor The S301 method is used to calculate the multidimensional features of the corridor and determine its traversability. If it is traversable, a reachability vector is generated. ,in The multidimensional magnitude of this vector; (2) Endpoint reachable vector : set of virtual endpoints points in As the endpoint, build from arrive straight corridor Similarly, a drivability determination is performed; if drivability is achieved, a reachability vector is generated. .

[0065] All accessible vectors constitute a tower position. The set of accessible vectors .

[0066] S303. Multidimensional Modulus Calculation

[0067] For any reachable vector Its multi-dimensional modulus The calculation formula is:

[0068] in Straight corridor area The diagonal length is used to normalize the horizontal distance. This makes the contributions of each dimension comparable; The elevation range of the corridor area is used to normalize the degree of elevation gradient change. The weight coefficients of each dimension are determined according to the AHP (Analytic Hierarchy Process), and in this embodiment, the values ​​are: (Horizontal distance weight) (Elevation fluctuation weight) (Obstacle weights) (Weight of cash crops). The principle for determining the weighting coefficient is: horizontal distance and terrain conditions are the primary considerations, obstacles are secondary, and cash crops are secondary factors.

[0069] Length The range of values ​​is The smaller the value, the lower the overall travel cost from the tower to the destination and the higher the path quality.

[0070] S304. Reachability Vector Filtering and Aggregation

[0071] S304-1. Vector Filtering: For each starting point The set of reachable vectors Calculate the minimum effective modulus Retain all that meet the requirements The reachable vectors form the filtered set. The basis for the screening ratio of 1.5 is to retain all candidate directions whose cost differs from the optimal path by less than 50%, thus ensuring the diversity of alternative paths while eliminating inferior solutions with excessively high costs.

[0072] S304-2. Global Aggregation: Merge the filter vectors of all starting points within a cluster to obtain a global vector set. .

[0073] S304-3. Node Extraction: Extraction The starting and ending coordinates of all vectors in the set form a node set. ,in This represents the total number of nodes. These nodes are the basic elements for subsequent closed-loop polygon detection.

[0074] S305. Closed-Loop Polygon Detection and Topology Analysis

[0075] S305-1. Adjacency Matrix Construction: For a set of nodes , build adjacency matrix ,in If and only if there exists a reachable vector that directly connects the node. and ,otherwise The adjacency matrix reflects the graph topology formed by the reachable vectors.

[0076] S305-2. Depth-First Search (DFS) for Loop Detection: A recursive DFS algorithm is used to traverse the entire graph and detect all polygons with loops. The specific execution flow is as follows: (1) Initialize the access tag array and the current path stack ; (2) For each unvisited node Perform a DFS traversal: ① will Mark as visited and push onto the stack; ② Traversal All adjacent nodes (Right now ),like If it is already in the stack, then retrieve it from the stack. The sequence of all nodes leading to the top of the stack is considered as a closed-loop polygon. ③ If If not visited, proceed recursively. DFS traversal; ④ During backtracking Pop it from the stack.

[0077] (3) Record all detected closed-loop polygons .

[0078] S305-3. Validity Filtering: Validity filtering is performed on the detected closed-loop polygons, and invalid polygons that meet any of the following conditions are removed: (1) The number of nodes is less than 3, i.e., a polygon cannot be formed; (2) The area of ​​the polygon is less than a preset area threshold. In this embodiment, m², to avoid pseudo-closed loops caused by excessively small geometric noise; (3) the polygons have self-intersections, which can be detected using the is_simple method of the Shapely library; (4) the polygons have duplicate vertices. The set of valid polygons after filtering is denoted as m². .

[0079] In this embodiment, the topology of group 5 obtained 1 valid closed-loop polygon after the above detection, group 12 obtained 2, and group 14 obtained 0.

[0080] S306. Virtual Starting Point Estimation

[0081] S306-1. Single polygon case: If If the number of elements is 1, then the effective polygon is defined as follows. The set of vertices is Calculate its geometric centroid:

[0082] S306-2. Case with multiple polygons: If If the number of elements is greater than 1 (e.g., group 12), calculate the geometric centroid for each polygon to obtain the centroid set. Set the centroids With virtual endpoint Merge to form a point set The Graham scan algorithm is used to calculate the convex hull (convex polygon) of the point set, and then the geometric centroid of the convex hull is calculated. The specific steps of the Graham scan algorithm are as follows: (1) Select the leftmost point in the point set as the reference point; (2) Sort the other points according to the polar angle; (3) Maintain a stack, traverse the sorted points, pop the points that are not turned left, and finally the remaining points in the stack form the vertices of the convex hull; (4) Calculate the centroid of the convex hull according to the centroid formula of the polygon.

[0083] S306-3. Case without polygons: If If the number of elements is 0, such as in group 14, where the tower positions are arranged in a straight line without forming a closed loop, then the geometric center of all tower positions within the group is taken as the centroid. .

[0084] S306-4. Terrain Adaptability Optimization: At the centroid around Within the neighborhood (neighborhood radius) Take 5 to 10 grid cells (corresponding to 17.5m to 35m), and calculate the passable grid cells for each cell. Overall rating The grid with the best overall score is used as the virtual starting point. .

[0085] The comprehensive scoring formula is:

[0086] The meaning and values ​​of each parameter are as follows: : grid The elevation value (unit: m) at the location is extracted from the DEM data; Center of mass around Average elevation (in meters) of all accessible grid cells in the neighborhood. Standard deviation of the elevation of all accessible grid cells within a neighborhood (in meters), reflecting the degree of elevation dispersion within the neighborhood; Obstacle indicator value, if grid If it is an obstacle, the value is 1; otherwise, it is 0. : Economic crop indicator value, if grid If it is a cash crop, the value is 1; otherwise, it is 0. : grid to the center of mass Euclidean distance (unit: m); : Elevation deviation penalty coefficient, in this embodiment, we take To control the sensitivity to elevation deviation; : Penalty coefficient for cash crops, taken in this embodiment Values ​​higher than This reflects the priority given to avoiding cash crops; Distance attenuation coefficient, taken in this embodiment as This causes the score to decrease gradually with increasing distance.

[0087] Overall score The meanings of each factor are as follows: (1) The first item is the elevation factor. The closer the grid elevation is to the average elevation of the neighborhood, the higher the score. Avoid setting the virtual starting point at the change in terrain. (2) The second item is the obstacle safety factor. Ensure that the virtual starting point is not located on obstacles. (3) The third item is the economic crop penalty factor. Try to avoid dense economic crop areas. (4) The fourth item is the distance decay factor. Prefer to choose a location close to the centroid.

[0088] Traversal Calculate the individual values ​​of all accessible grid cells within the neighborhood. The value is used as the virtual starting point, and the grid cell corresponding to the maximum value is selected. .

[0089] In this embodiment, the virtual starting point calculation process for group 5 is as follows: Figure 3 As shown, the centroid After terrain-adaptive optimization, it shifted approximately 12 meters southeast, avoiding local depressions and ponds; the virtual starting point of cluster 12 is as follows: Figure 4 As shown, due to the corners and dense road distribution within the cluster, the virtual starting point, after two rounds of closed-loop polygon calculations, is selected at a gentle point between T21, T22, and the road network. The virtual starting point for cluster 14 is shown below. Figure 5 As shown, since the tower positions are arranged linearly and do not form a closed loop, the virtual starting point is directly selected at a flat elevation near tower position T26.

[0090] S4. Improvement A The algorithm performs main path planning.

[0091] This step uses an improved A algorithm, starting from a virtual starting point. To existing road collection Search for the optimal backbone path. Improve the A algorithm on the standard A algorithm. Based on this, a three-level constraint verification mechanism and a multi-terminal heuristic function have been added.

[0092] Algorithm data structure definition

[0093] Define A The data structure for the search node is as follows: Current node : The coordinates of the raster node currently expanding in the search space; Parent node : Expand to the coordinates of the parent node of the current node; starting point The starting point of the main path for the current group; Virtual endpoint set Candidate endpoints for the main route; Existing road collection : All existing road grid points, as the actual endpoint set; Value: The actual cost of moving from the starting point to the current node; Value: Heuristically estimated cost from the current node to the destination; value: , used for priority queue sorting.

[0094] S401. Constraint-Aware Node State Extension and Verification Mechanism

[0095] An 8-neighbor expansion strategy (i.e., expanding from the current node to its adjacent grids in the eight directions of east, south, west, north, southeast, northeast, southwest, and northwest) is adopted to generate 8 candidate neighbor nodes for each current node, and a three-level constraint verification is performed on each candidate node: S401-1. Level 1 Obstacle Avoidance Verification: Check candidate nodes Is the attribute of an obstacle? (Including buildings, bodies of water, steep cliffs, etc.), the inspection method is: query in the obstacle raster layer. The corresponding grid's encoding value, if the value is an obstacle code (in this embodiment, the house code is 2, the water body code is 3, and the cliff code is 4), is determined to be an obstacle, and the candidate node is directly removed and does not participate in subsequent verification.

[0096] S401-2. Second-level slope verification: Calculate the current node. to candidate node Slope value:

[0097] in The elevation difference (in meters) between the candidate node and the current node. m is the grid width, i.e., the horizontal distance the road extends. When If the slope of a candidate node exceeds the climbing ability of the construction machinery, it is directly eliminated.

[0098] S401-3. Level 3 Turning Radius Verification: Based on Parent Node Current node and candidate nodes The three-node sequence is used to calculate the turning radius. : (1) Calculation arrive Horizontal projection vector:

[0099] (2) Calculation arrive Horizontal projection vector:

[0100] (3) Calculate the angle of change of direction :

[0101] in:

[0102] (4) Calculate the turning radius :

[0103] in .

[0104] like If m is used to determine that the turning radius of the candidate node does not meet the minimum turning requirements of the construction machinery, it will be directly eliminated.

[0105] Candidate nodes that pass the three-level constraint verification enter the subsequent cost calculation and priority queue, while those that fail are recorded in the pruning record table.

[0106] S402. Multi-terminal heuristic function

[0107] S402-1: Before performing complex physical cost calculations, perform a quick topological connectivity check: (1) Get the current node Connectivity label The connected component labels are calculated using the two-pass algorithm: the first scan assigns a temporary label to each grid cell and records the equivalence relation; the second scan merges the equivalence labels, and finally each connected component has a unique label value.

[0108] (2) Traverse all virtual endpoints , obtain Connectivity label .like This indicates and If the virtual destination is not within the same passable area, it is directly determined to be unreachable, and the candidate destination set is selected. Remove from the list.

[0109] (3) Secondary filtering based on S302 reachability calculation method: For virtual endpoints filtered by connected components, construct arrive Calculate obstacle density in a straight corridor. ,like If the corridor is deemed impassable, the virtual endpoint will also be determined from... Remove from the list.

[0110] After the above two layers of screening, the effective endpoint set is obtained. ,in .

[0111] S402-2. Dynamic Heuristic Function Calculation: Traversing the Valid Endpoint Set Each endpoint in Calculate the composite heuristic value: (1) Basic movement cost : The three-dimensional Euclidean distance from the current node to the destination.

[0112] (2) Nonlinear terrain correction : in For terrain correction factors, this embodiment takes... ; The actual slope of the straight line from the current node to the endpoint is calculated using the following formula: ; For the optimal slope, this embodiment takes... (Economical climbing angle for tracked machinery). This correction term results in smaller heuristic values ​​for paths with gradients close to the optimal gradient, guiding the search towards gentler terrain.

[0113] (3) Economic penalties Penalties are imposed based on the proportion of cash crops within the straight-line corridor from the current node to the destination.

[0114] in The cost coefficient for economic crops is taken as [value missing] in this embodiment. ; From arrive The number of economic crop grids within the straight corridor area; This represents the total number of grid cells within the corridor area. This penalty causes the path to actively avoid areas densely populated with cash crops.

[0115] (4) Detour penalty :

[0116] in From arrive The number of obstacle grids within a straight corridor area. The higher the proportion of obstacles, the greater the expected cost of detours.

[0117] (5) Composite heuristic value:

[0118] S402-3. Multi-terminal heuristic value synthesis: Take the minimum composite heuristic value among all valid endpoints as the heuristic value of the current node.

[0119] This design enables A The search always expands towards the virtual endpoint that is "nearest and has the lowest overall cost," avoiding suboptimal paths due to a fixed single endpoint.

[0120] S403.A Search execution and termination strategy

[0121] S403-1. Initialization: Set the virtual starting point As the starting node, it is added to the priority queue (the priority queue is implemented using Python's heapq module, and is ordered by...). (Values ​​sorted in ascending order), settings ,calculate , Initialize a hash table (Python dict) of visited nodes to store visited nodes and their current best position. Values ​​should be set to avoid redundant expansion.

[0122] S403-2. Main Loop: Executes the following sub-steps repeatedly until the termination condition is met: (1) Pop from priority queue The node with the smallest value is selected as the current node. ; (2) If The attribute belongs to the existing road set (Right now If the existing road network is already connected, then record the total cost of the current path. Set global cost threshold (This embodiment takes) This means that subsequent searches are allowed to find alternative paths that are no more than 10% more expensive than the current optimal path, and the state of finding the destination for the first time is marked, but the search does not terminate immediately, and continues to find possible better paths or equivalent alternative paths; (3) Generate a set of neighboring nodes whose constraints have been verified for the current node (according to the three-level constraint verification process of S401). (4) For each verified neighbor node : like Unvisited or new path cost Smaller than the recorded Then update Information: Recalculate and ,Will Add to the priority queue and the visited table; in From arrive The movement cost is calculated by taking the three-dimensional Euclidean distance between the two nodes; (5) Check termination conditions: If the priority queue is empty (all reachable nodes have been expanded) or the currently popped node is empty. Value exceeds cost threshold If the cost of all remaining candidate paths exceeds 110% of the current optimal path, then the search is terminated.

[0123] S403-3. Optimal Path Extraction: Select the node with the minimum value from the recorded endpoint nodes. The node with the value is the optimal endpoint. By tracing back the parent node chain from the visited table, the complete set of optimal path grids can be reconstructed. ,in , .

[0124] S404. Grid Map Update

[0125] Main path All grid nodes (except the start point) (External) Mark these grid cells on the map as "End of Built Road", that is, update the attributes of these grid cells to be passable roads, and use them as the current group. Collection of the ends of existing roads Initial element: These grids will serve as accessible target points in subsequent branch route planning.

[0126] In this embodiment, the main path The total length is 340m, the average slope is 12°, and the minimum turning radius is 7.2m, which meets the constraints of construction machinery.

[0127] S5. Tower Accessibility Ranking

[0128] This step is for group formation. Each tower location in the calculation is located from its virtual starting point. The reachable vectors are sorted by magnitude from smallest to largest to determine the order of branch path planning. The purpose of this sorting is to prioritize planning tower locations closer to the virtual starting point, so that their branches can be built and connected to the endpoint set as early as possible. This provides more "intermediate access points" for subsequent tower sites located further away, maximizing the benefits of road reuse.

[0129] S501. Reachability Vector Construction: For Each tower position ,by Starting point, virtual starting point Construct a straight corridor area with the destination as the endpoint. The obstacle density of the corridor was calculated using the S301 method. Density of cash crops Horizontal distance and degree of elevation gradient change .

[0130] S502. Roadworthiness determination: If Then determine arrive If the passage is blocked, the tower location will be marked as "redirection required" and processed separately; if the passage is permitted, the multi-dimensional modulus length will be calculated according to the S303 formula. .

[0131] S503. Sorting: Sort all accessible tower positions by Sort the tower positions in ascending order to obtain the sorted tower position sequence. ,in The tower position with the smallest module length (the nearest tower position). The tower position with the largest module length (the farthest tower position).

[0132] In this embodiment, the modulus lengths of the three tower positions T26, T27, and T28 in cluster 5 are 215, 241, and 326 respectively, and the sorting result is as follows: .

[0133] S6. Branch Route Planning

[0134] This step is based on the sorted tower position sequence. Plan the branch paths for each tower location sequentially, and during the planning process, determine the endpoint set. Continuous dynamic growth enables iterative reuse of "existing roads".

[0135] S601. Initialization: Set the current tower position index. Tower location path set Current destination set The initial trunk path is generated in S4. All grid points (excluding the starting point) ).

[0136] S602. Individual tower site branch line planning: (1) Take the current tower position As the starting point of the path; (2) Using the current endpoint set As a multi-endpoint target set (Note: Not only includes the trunk path It also includes the branch path grid of all planned preceding tower sites. (3) Adopt the improved A described in S4 Algorithm (including three-level constraint verification and multi-terminal heuristic function), search from arrive Optimal path ; (4) During the search process, the set of destinations Any grid cell in the search can be used as a termination condition; that is, once a node is found to belong to a grid cell, the search is terminated. This means that the current tower location is considered to have successfully connected to the existing road network, and the path is recorded.

[0137] S603. Dynamic updates of grid maps: (1) The newly planned branch routes All grid nodes (excluding the start point) Marked as "End of Existing Road"; (2) Add these grids to the endpoint set. : ; (3) Update the global grid map, Modify the properties of the middle grid to allow passageways.

[0138] S604. Iterative Control: Let... ,like ( for If the total number of tower positions is zero, return to S6-2 to continue planning the next tower position; otherwise, complete the branch path planning for all tower positions in the current cluster.

[0139] In this embodiment, the branch line planning process for group 5 is as follows: Figure 6 As shown: T6 was planned first, and its branches directly connect to the main route. (Red), the access point is located at The mileage marker is K0+380; during the T7 planning, the endpoint assembly point was... Not only includes It also includes the T6 branch. The algorithm automatically detects that connecting the T7 branch to the middle section of the T6 branch is better than directly connecting it. The design is superior (total length reduced by approximately 65m, and gradient gentler), achieving "secondary sharing" of branch lines; during the planning of T8, the end of branch line T7 was further connected. The three branch lines are connected in series and then connected to the main trunk via branch line T6, reducing the total length by approximately 18% compared to three branch lines independently connected to the main trunk. The final route planning results for Cluster 5 are as follows: Figure 7 As shown.

[0140] S7. Cross-group collaboration and global path inheritance

[0141] S701. Inter-group inheritance mechanism: Traverse all groups For each group, the complete process from S3 to S6 is executed sequentially. Upon entering the next group, the previous group's... As the initial element, it is inherited into the current group's endpoint set.

[0142] Specifically, for the first group... Its final set Includes only the trunk path Grid points (initialized by S4); for clusters ( Before executing the S4 trunk path planning, the inheritance operation is performed first: This means that the initial set of endpoints for the current cluster includes all the endpoints of the existing roads in the previous cluster, as well as the main road of the current cluster itself. Thus, in stage S6, the branch lines of the current cluster can not only directly connect to the main road of its own cluster, but also to the roads built in the previous cluster, enabling cross-cluster road sharing.

[0143] S702. Path optimization effect across clusters: In this embodiment, cluster 5 and cluster 6 are adjacent (tower spacing is approximately 1.2km). After the planning of cluster 5 is completed, its existing road network... Includes trunk path And three branch lines, with a total length of approximately 2.1km. When entering the planning phase of Group 6, Inherited For all road nodes, when planning the branch line for tower T29 in Group 6, the algorithm search found that its optimal access point is located at... Near the end of the T27 branch line (rather than the main road of this cluster) Only about 230m of new branch line needs to be built to connect to the existing road network, which saves about 65% of the branch line length compared to independent access to the existing road (which requires the construction of about 650m).

[0144] S703. Termination Condition: After traversing all 16 clusters, the paths of all tower locations have been connected to the global road network, and the algorithm terminates. Output all path planning results, including the grid coordinate sequence of each path segment, path length, maximum slope, minimum turning radius, area crossed by economic crops, obstacle avoidance statistics, etc.

[0145] The embodiments described herein are preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Therefore, all equivalent changes made in accordance with the structure, shape, and principle of the present invention should be covered within the scope of protection of the present invention.

Claims

1. A method for planning temporary access roads for transmission lines in a multi-tower collaborative planning approach, characterized in that, Includes the following steps: S1. Acquire data on terrain features, sensitive areas, tower coordinates, and DEM elevation within the power transmission corridor, and set a minimum permissible turning radius threshold. Maximum allowable slope and road width And construct a grid map and determine the tower clusters. At the same time, construct a set of grid points for existing roads. ; S2. To Each group ,by The predetermined range is extended to the front, back, left, and right sides of the inner tower location. The inner tower raster map is extracted from the raster map, and the virtual endpoint in the inner tower raster map is obtained using a clustering analysis-based method. ; S3. To Each group The virtual starting point is obtained by using a grid corridor analysis and multi-dimensional vector aggregation method. ; S4. Adopting improved A Algorithm obtained to Optimal path grid set And update the raster map, The middle grid update is the endpoint, with All grid points and All elements in the current group Collection of the ends of existing roads The initial element; S5. Calculation Each tower to Find the reachable vector magnitudes and sort them from low to high to obtain the sorted vectors. ; S6. From At the beginning, for Repeat the following steps for all elements: Starting from the current set of existing road endpoints on the raster map. As the endpoint, adopt improved A The algorithm obtains the paths to all tower locations. And update the raster map, Updated to endpoint and added ; S7. Traversal Each group within the group repeats S2-S6, and upon entering the next group, the previous group's... As the initial element, it is inherited into the current group's endpoint set until completion. Search all tower location paths.

2. The method for planning temporary access roads for transmission lines based on multi-tower collaborative planning according to claim 1, characterized in that, S2 includes the following steps: S201. Three-dimensional feature vector construction: Extracting the existing road set from the inner tower grid map. Intersecting grids serve as inner tower road grids. Each inner tower road grid is represented as a point in three-dimensional space, and its feature vector is... ; S202. Perform Z-score standardization so that the mean of each dimension is 0 and the standard deviation is 1; S203. Perform 3D DBSCAN clustering to obtain cluster labels and intra-cluster road points. ; S204. Generate a virtual endpoint: (This is related to...) For each cluster, calculate the feature vector of all road points within it. The arithmetic mean of each dimension is used as the three-dimensional centroid coordinates of the cluster, serving as the virtual endpoint. .

3. The method for planning temporary access roads for transmission lines based on multi-tower collaborative planning according to claim 1, characterized in that, S3 includes the following steps: S301. To Multidimensional feature extraction was performed on each tower location: Constructing tower sites To the finish line Straight corridor area This region contains all lines. left and right sides The distance of each grid cell, where The value is a straight line Length ; Calculate horizontal distance ; Calculate obstacle density ; Calculate the density of cash crops ; Calculate the mean absolute elevation difference ; Calculate the maximum elevation difference ; Calculate the rate of change of elevation ; Calculate the degree of change in elevation gradient ,in, refer to The number of grid cells, Refers to obstacle grids, Refers to the grid of economic crops. This refers to the elevation difference between adjacent grid cells. This is an indicator function; it takes a value of 1 when two adjacent elevation changes are in opposite directions, and 0 otherwise. and These represent the maximum and minimum elevation values ​​within the region, respectively. These are the weighting coefficients; like ,in If a preset obstacle threshold is not met, the corridor is determined to be impassable, and the vector is invalid. S302. Construction of Multi-Dimensional Reachable Vectors: [The sentence is incomplete and likely refers to a separate topic Other tower locations within the cluster Construct reachable vectors for adjacent tower locations to the endpoint. , with virtual endpoint set Construct a destination reachable vector from each point in the vector. Based on the corridor analysis method of S301, the traversability of each vector is determined one by one, and a set is formed from all traversable vectors. ; S303. Calculate multidimensional modulus For any vector The formula for calculating its modulus is: , in The length of the diagonal of the straight corridor area. , , , ; S304. Reachability Vector Filtering and Aggregation: For each starting point Calculate the minimum effective modulus Retain all that meet the requirements The reachable vectors form the filtered set. Merging the filtering vectors from all starting points yields the global vector set. Extract the start and end coordinates of all vectors to form a node set. ; S305. Closed-Loop Polygon Detection and Topology Analysis: For Node Sets Construct an adjacency matrix, use depth-first search to detect closed polygons, and filter out polygons with areas smaller than a preset threshold, or those with self-intersections or duplicate vertices to obtain an effective set. ; S306. Virtual Starting Point Estimation: like Calculate the geometric centroid of an element with a quantity of 1. ,in The number of vertices; like If the number of elements is greater than 1, calculate the geometric centroid of each polygon, merge the set of centroids with the virtual endpoint, and use the Graham scan algorithm to calculate the convex polygon and find its centroid. ; At the center of mass around Within the neighborhood, calculate each passable grid cell. Overall rating , in For elevation, The average elevation of the neighborhood. For elevation standard deviation, Indicator values ​​for cash crops, The distance from the grid to the centroid. These are preset coefficients; The grid with the best overall score is used as the virtual starting point. .

4. The method for planning temporary access roads for transmission lines based on multi-tower collaborative planning according to claim 1, characterized in that, Improvement A in S4 The algorithm includes a constraint-aware node state expansion and verification mechanism, and the steps are as follows: S401. An 8-neighborhood expansion strategy is used to generate candidate neighbor nodes for each current node, and three-level constraint verification is implemented: S401-1. Level 1 Obstacle Avoidance Verification: Check if the candidate node attribute is an obstacle; if so, remove it. S401-2. Second-level slope verification: Calculate the current node. to candidate node slope value ,in The elevation difference between adjacent grid cells. The grid width (i.e., road width) is... Then remove; S401-3. Level 3 Turning Radius Verification: Based on Parent Node Current node and candidate nodes The three-node sequence is used to calculate the turning radius. : Calculate the horizontal projection vector ， ， Calculate the angle of change of direction , in , , ； Calculate the turning radius ,in ,like Then remove it.

5. A method for planning temporary access roads for transmission lines based on multi-tower collaborative planning according to claim 4, characterized in that, Improvement A in S4 The algorithm also includes the following steps: S402. Multi-terminal heuristic function, the specific steps are as follows: S402-1. Topological reachability pre-screening of virtual endpoints: Obtain the connected component labels of the current node, remove virtual endpoints with inconsistent connected component labels, and perform a secondary screening based on the reachability calculation method in S302 to obtain the set of valid endpoints. ; S402-2. Dynamic Heuristic Function Calculation: Traversal The elements in the array, for each endpoint Calculate the basic moving cost Nonlinear terrain correction ,in This is the terrain correction factor. The actual slope from the starting point to the end point. The optimal slope; Economic penalties ， in For the grid cost coefficient of cash crops, This represents the number of economic crop grid cells within a straight corridor. This represents the total number of grid cells in the corridor. Detour penalty ,in Number of obstacle grids; composite heuristic value ; S402-3. Heuristic Value Composition and Guidance for Multiple Endpoints: Take the minimum composite heuristic value among all valid endpoints as the heuristic value of the current node. 。 6. The method for planning temporary access roads for transmission lines based on multi-tower collaborative planning according to claim 1, characterized in that, Improvement A in S6 The algorithm uses the sorted tower position sequence The order is based on the current set of completed road endpoints. For multi-destination path search, after completing the path planning for each tower location, all grid nodes of that path are added. This allows the endpoint set to grow dynamically.

Images