High mountain wind power plant road gradient and curve optimization design system and method

Abstract

The invention discloses a high mountain wind power plant road gradient and curve optimization design system and method. A digital terrain grid and topology analysis technology is utilized to realize automatic generation and optimization of a road alignment. A digital terrain model is established according to elevation, gradient and landform data, and extreme points, saddle points and a connection relation are extracted through Morse-Smal topological decomposition to form a topological skeleton line. And generating a slope direction consistency vector field, and obtaining a primary road alignment set. Schwarz-Christoffel shape-preserving mapping is performed on the local terrain region, a two-dimensional parameter domain terrain model and a positive and negative mapping relation are constructed, double-domain constraint screening of a topological domain and a parameter domain is realized, and curvature or gradient abnormal paths are eliminated. And generating a path line shape under the original terrain coordinates through inverse mapping reconstruction, and extracting the space length, the segmented gradient and the turning radius of the path to form a road geometric structure data set. According to the method, automatic optimization and geometric parametric modeling of the road alignment under the complex terrain condition are realized.

Description

Technical Field

[0001] This invention relates to the fields of road engineering and wind farm construction technology, and in particular to a system and method for optimizing the design of road slope and curves in high-altitude wind farms. Background Technology

[0002] With the increasing development of wind power resources in mountainous and complex terrain areas, the planning and design of internal roads in wind farms has become a crucial factor restricting construction cycles and operation and maintenance efficiency. Currently, the design of roads in high-altitude wind farms largely relies on manual route selection and traditional terrain analysis software. The road alignment is mainly determined by manually comparing elevation data overlaid on two-dimensional topographic maps. This method has significant limitations in mountainous environments with large terrain undulations and drastic slope changes: on the one hand, traditional digital elevation models can only provide macroscopic slope information and cannot accurately describe key geomorphic features such as local saddle points and extreme points, resulting in insufficient response to micro-topography and problems such as excessive slope or excessively small turning radii; on the other hand, commonly used path planning algorithms such as A* and Dijkstra's fail to comprehensively consider slope constraints and curvature continuity. Although the generated paths are computationally feasible, they are often not feasible in engineering construction, requiring multiple manual corrections, increasing design costs and timelines. In addition, existing road optimization models are mostly based on single-domain geometric parameters and lack a dual-domain constraint mechanism of topological domain and parameter domain, making it difficult to achieve a balance between the continuity and smoothness of the path under complex terrain.

[0003] Therefore, how to provide a system and method for optimizing the design of road slope and curves in high-altitude wind farms is a problem that urgently needs to be solved by those skilled in the art. Summary of the Invention

[0004] One objective of this invention is to propose a road slope and curve optimization design system and method for high-altitude wind farms. It utilizes terrain intelligent analysis and dual-domain mapping technology that combines Morse-Smale topology decomposition and Schwarz-Christoffel conformal mapping to achieve automatic generation and optimization of road paths under complex terrain.

[0005] A road slope and curve optimization design system and method for high-altitude wind farms according to an embodiment of the present invention includes the following steps: The terrain modeling module takes regional elevation raster data, slope field data, and geomorphic feature data as input and generates a digital terrain grid. The topology deconstruction module receives the digital terrain mesh, performs Morse-Smale topology decomposition, extracts extreme points, saddle points and their connection structures, and constructs a set of topology skeleton lines. The orientation field generation module constructs a slope-aspect consistent vector field based on the topological skeleton line set, generating a primary road alignment set; The mapping construction module selects a local grid region containing the primary road alignment set, performs Schwarz–Christoffel conformal mapping, generates a two-dimensional parametric domain terrain model, and simultaneously generates a set of forward and inverse mapping relationships; The path filtering module receives a set of primary road alignments, a two-dimensional parameter domain terrain model, and a set of forward and inverse mapping relationships. It eliminates disconnected paths in the topological domain and filters out paths with excessive curvature or slope in the parameter domain, generating a path candidate strip that satisfies the dual-domain constraints. The reflection reconstruction module receives the path candidate strip and the set of forward and inverse mapping relationships, and transforms the path candidate strip to the original digital terrain grid coordinate system through the inverse Schwarz-Christoffel mapping to reconstruct the corresponding path shape. The parameter extraction module extracts the spatial projection length, segmented slope changes, and turning radius sequences based on the path alignment to construct a road geometry dataset.

[0006] Optionally, modules can be integrated using the following methods: A digital terrain grid is generated based on the elevation raster data, slope field data, and geomorphic feature data of the target wind farm area. Morse-Smale topological decomposition is performed on the digital terrain mesh to extract extreme points, saddle points and their connection relationships, and generate a set of topological skeleton lines. The local aspect consistency vector field is calculated based on the set of topological skeleton lines to generate a set of primary road alignments that express the natural travel direction; For a local terrain grid region covering the primary road alignment set, perform Schwarz–Christoffel conformal mapping to obtain a two-dimensional parametric domain terrain model, and simultaneously generate a set of forward and inverse mapping relationships; In the topology domain, primary lines that are not connected to the skeleton structure are removed. In the parameter domain, lines with abrupt curvature changes or slopes exceeding the threshold are identified and excluded. Based on the results of the two domains, path candidate zones are generated. By using the inverse mapping relationship, the path candidate band is mapped from the parameter domain back to the original digital terrain grid coordinate system, generating the path shape corresponding to the terrain; Based on the path alignment, the spatial projection length, segmented slope changes, and turning radius sequence are calculated to construct a geometric structure dataset for road design.

[0007] Optionally, generating the set of topology skeleton lines includes: Using a digital terrain grid as input, the elevation value of each grid cell is written into a scalar field array, and point-by-point gradient calculation is performed on the scalar field array, recording the gradient direction and gradient magnitude of each grid cell; For each grid cell in the scalar field array, compare its elevation relationship with that of its neighboring cells point by point. Grid cells with elevations higher than all neighboring cells are marked as maxima, those with elevations lower than all neighboring cells are marked as minima, and grid cells with abrupt changes in gradient direction and reverse changes in gradient magnitude are marked as saddle points, thus forming a list of key point indices. Input the list of key point indices into Morse-Smale topology decomposition, classify the key point numbers, and output a key point type table and a topology output table showing the connection relationships between key points. Read the key point number pairs recorded in the topology output table, perform line segment interpolation between each pair of key points according to the coordinate position of the number pairs in the digital terrain grid, and write the interpolated line segment sequence into the trajectory line list in spatial order. The list of trajectory lines is checked for breakpoints. Line segments that do not meet the continuity requirements are filled with coordinate points to ensure that each trajectory line maintains the same spatial resolution as the digital terrain grid. Duplicate segments are cleaned up and the sequence numbers are sorted for all trajectory lines. Establish a correspondence between the maximum point number, minimum point number, and saddle point number in the key point type table and the line segment sequence in the trajectory line list, and bind each trajectory line to its start and end key point number. The key point coordinates, key point numbers, key point types, and their corresponding trajectory line sequences together constitute a set of topological skeleton lines.

[0008] Optionally, generating the initial road alignment set includes: Using the set of topological skeleton lines as input, the grid index sequence of each skeleton line is read, and the skeleton line coverage area is marked in the digital terrain grid. The skeleton line unit and its neighboring unit are written into the candidate region table. In the candidate region table, the slope aspect vector is calculated for each cell, where the slope aspect vector is the vector pointing from the current grid cell to the adjacent cell in the direction of maximum slope, and all slope aspect vectors are recorded in the slope aspect vector table; In the aspect vector table, the aspect vectors belonging to the skeleton line unit are assigned a direction number. The direction number is written into the skeleton direction table along with the grid index. The consistency of the direction number order of adjacent units in the skeleton direction table is checked. Units whose direction number changes more than a preset threshold are registered as direction transition units. Within the candidate region table, starting from each skeleton line unit, advance along its corresponding direction by numbering to construct a unit advancement sequence. Write the index of each unit in the advancement sequence into the direction advancement table. When advancing to a direction transition unit, read the slope vector of its adjacent unit, and select the unit with the smallest angle among the adjacent units as the next advancement unit based on the angle relationship between the slope vector and the current advancement direction. Write it into the direction advancement table and continue advancing. The propulsion sequences in all directions of the propulsion table are traversed. The grid index continuity is checked in each propulsion sequence. Linear interpolation is performed at the positions where index jumps occur. The interpolated cells are written into the propulsion sequence to keep the propulsion sequence continuous in the grid coordinate space. In each continuous advance sequence, record the starting skeleton line element index, the ending element index, and all the grid element indices contained therein. Define the sequence as a directional travel line element and write all directional travel line elements into the directional line element table. In the directional alignment table, the slope aspect vectors of all alignment units are checked for monotonicity. Alignment units whose slope aspect vector sequences satisfy the monotonicity condition are marked as primary road alignments, and alignment units that do not satisfy the monotonicity condition are deleted from the directional alignment table. All linear elements that pass the monotonicity check are combined according to their start and end element indices, element sequence, and direction number to form a primary road alignment set.

[0009] Optionally, generating a set of forward and reverse mapping relationships includes: Using the primary road alignment set as input, the digital terrain grid index range containing all primary alignments is extracted, the corresponding local terrain region is located in the digital terrain grid, and the elevation values ​​of all grid cells in the region are read to form a local terrain data matrix. In the local terrain data matrix, a linear region containing the direction of maximum slope is selected as the reference baseline. Grid-weighted smoothing is performed on the elevation data on both sides of the reference baseline to generate a set of boundary isopleths, and the set of isopleths is used as the input of local boundary data. Perform corner detection on local boundary data to determine the boundary inflection point index sequence. and the included angle between adjacent boundary segments Piecewise discretization is performed to decompose the continuous curve boundary into a finite piecewise line segment structure; Insert auxiliary nodes within each boundary polyline segment, ensuring that the distance between adjacent nodes within each segment is less than a preset boundary resolution threshold. And calculate the local normal direction vector of each node. ; Based on the angle between the normal direction of the boundary node and the local slope direction Establish node weighting functions , weight Write to the boundary weighting table; In the boundary weighted table, all nodes are sorted by weight. Sort the nodes from largest to smallest, select a fixed proportion of the node set to form the boundary skeleton node set, and connect them in the order of this node set to form the input boundary vertex sequence of the conformal mapping. The angle between adjacent line segments of the boundary vertex sequence Write the angle parameter table and perform normalization; Using the boundary vertex sequence and angle parameter table as input, construct a complex plane boundary vertex array. ,in Represented by complex coordinates; Define the integral kernel function for the Schwarz–Christoffel conformal mapping; Establish a mapping integral solution grid, and assign coordinates of each node within a subset of the local terrain grid. Mapped to complex plane coordinates And perform integration on each node: Gaussian quadrature is applied to the boundary nodes; The trapezoidal integral method is used for internal nodes; The real and imaginary parts of the integral result are represented as two-dimensional parameter domain coordinates. ; All nodes in the parameter field Coordinates and their corresponding original grid index numbers Write the parameters to the parameter field node table and index them in spatial order; Divide the parameter domain node table into Each parameter block contains a fixed number of nodes; Within each parameter block, a node index mapping table is established, using the original mesh index number as the key and the parameter domain coordinates as the key. As the value, write it to the forward mapping sub-table, and mark the sub-table for each parameter block as... ; After generating all forward mapping sub-tables, for each parameter block Perform reverse matching and construction operations: Read sub-table All key-value pairs in the database; Using parameter domain coordinates Using the key and the original index number as the value, a reverse mapping sub-table is generated. ; Perform two-dimensional Delaunay triangulation on the inverse coordinate keys within each parameter block, and establish a coordinate interpolation function based on the triangulation results. ; Perform overlapping boundary detection on all parameter blocks. When duplicate mapping points are detected at the boundary of a parameter block, retain the key-value pair with the smallest mapping error and delete the rest of the mapping records to form a unified forward mapping table and reverse mapping table. In the global forward mapping table, the mapped coordinate data is standardized to floating point, and the parameter field coordinates are retained to four significant decimal places. In the global inverse mapping table, for all inverse interpolation functions The output is linearly normalized once to ensure that the mapped coordinate values ​​are at the same scale as the original grid space. After completion, the forward mapping table, reverse mapping table and interpolation function set are merged to form a mapping relationship set, and its parameter domain boundary index, number of nodes and coordinate distribution range are recorded; Within a subset of the local terrain grid, N pairs of mapping verification points are randomly selected. For each verification point, the angle preservation error after mapping is calculated. Error compared to side length When satisfied and When this happens, the mapping set is marked as valid. and To set a threshold; Index compression is performed on the effective mapping relationship set, merging the forward and reverse mapping tables into a unified index structure, and recording the parameter domain node coordinates, boundary vertex sequences, and angle parameter tables in the same data block; The above unified index structure is named the two-dimensional parameter domain terrain model.

[0010] Optionally, the generated path candidate bands include: Using the primary road alignment set as input, the original terrain grid index sequence contained in each alignment is read one by one, and the skeleton line number that has a common grid cell with the index sequence is retrieved in the topological skeleton line set; When there are no grid cells in the index sequence of a primary line shape that coincide with the set of topological skeleton lines, the line shape is marked as a topologically disconnected line shape and removed from the set of primary road line shapes, resulting in a set of line shapes that retain topological constraints. For each line shape in the set of line shapes that are retained for topological constraints, the line shape index sequence is mapped to the two-dimensional parameter domain coordinate system according to its grid index order, generating a parameter domain coordinate sequence that corresponds one-to-one with the line shape, and the parameter domain coordinate sequence is written into the parameter line shape table; In the parameter line table, for each parameter line, extract the coordinate points of three adjacent parameter domains in sequence, construct a list of polyline segment units segment by segment, and record the turning direction number of each polyline segment. For each parameter line shape, perform sequential traversal of the polyline segment cell list. When the turning direction number of a continuous polyline segment changes in the opposite direction, record the corresponding polyline segment position as a geometric anomalous node and mark the line shape containing the node as a direction change line shape. For each parameter line shape, calculate the ratio of the longitudinal change to the lateral distance between the coordinates of adjacent parameter domains according to their coordinate order, and write the ratio into the slope sequence. When more than a preset number of abnormal ratios appear consecutively in the slope sequence, the corresponding line shape is marked as an abnormal slope line shape. Parametric lines marked as abrupt direction changes or anomalous slopes are removed from the parametric lines table, forming a set of parametric domain constraint-preserved lines. Perform intersection matching between the topology constraint-preserved line set and the parameter domain constraint-preserved line set according to the line set number, retain only the line sets that exist in both sets, and generate a bi-domain consistent line set; For the linear shapes in the dual-domain consistent linear shape set, the spatial distance between adjacent linear shapes is expanded according to their index sequence in the original terrain grid, and multiple linear shapes located on both sides of the same topological skeleton line and with a spacing less than the preset width are merged into a continuous strip structure. Write the start and end indices of each strip structure, the set of line numbering contained therein, and the corresponding grid index range into the path candidate table.

[0011] Optionally, generating path alignments corresponding to the terrain includes: based on the coordinate index sequence of each alignment in the parameter domain within the candidate path zone, using the inverse mapping relationship of the corresponding parameter blocks, mapping the path points in the parameter domain back to the original digital terrain grid coordinate system through centroid interpolation under Delaunay triangulation, performing spacing supplementation and elevation interpolation on the mapped path points, constructing a continuous three-dimensional path point sequence according to the original parameter domain order, and organizing the set of polyline segments after continuity correction into a path alignment set dataset containing three-dimensional coordinate information according to the path number.

[0012] Optionally, constructing a geometric structure dataset for road design includes: based on the three-dimensional coordinate sequence of each path in the path alignment set, sequentially calculating and recording the cumulative spatial projection length of each path, the segmented slope change value between adjacent nodes, and the turning radius sequence determined by three consecutive nodes, and storing the spatial projection length, segmented slope change, and turning radius sequence in association according to the path number to form a geometric structure dataset for road design.

[0013] The beneficial effects of this invention are: This invention introduces the Morse–Smale topology decomposition algorithm and the Schwarz–Christoffel conformal mapping model to achieve topological structuring and geometric continuity control for road alignment in complex mountainous terrain. This combined method can automatically identify terrain extrema, saddle points, and their connections, construct a set of terrain skeleton lines, and perform dual-domain mapping filtering within the parameter domain. This effectively avoids the local optima and abrupt slope changes present in traditional two-dimensional terrain alignment, thus ensuring smooth continuity of road slope and curvature.

[0014] This invention establishes a dual-domain constraint mechanism involving both the topological and parameter domains, integrating slope field, curvature threshold, and topological connectivity conditions to perform multi-level screening and optimization of the initial road alignment, significantly improving the structural rationality and geometric constructability of the path. Compared to single-domain optimization models, this method maintains consistency in path continuity and terrain adaptability under complex terrain conditions, reducing the need for subsequent manual corrections. Attached Figure Description

[0015] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1 This is a flowchart of a road slope and curve optimization design system and method for high-altitude wind farms proposed in this invention; Figure 2 This is a flowchart illustrating the construction process of a two-dimensional parametric domain terrain model based on Schwarz–Christoffel conformal mapping proposed in this invention. Detailed Implementation

[0016] The present invention will now be described in further detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0017] refer to Figures 1-2 A system and method for optimizing the design of road slope and curves in high-altitude wind farms, comprising the following steps: The terrain modeling module takes regional elevation raster data, slope field data, and geomorphic feature data as input and generates a digital terrain grid. The topology deconstruction module receives the digital terrain mesh, performs Morse-Smale topology decomposition, extracts extreme points, saddle points and their connection structures, and constructs a set of topology skeleton lines. The orientation field generation module constructs a slope-aspect consistent vector field based on the topological skeleton line set, generating a primary road alignment set; The mapping construction module selects a local grid region containing the primary road alignment set, performs Schwarz–Christoffel conformal mapping, generates a two-dimensional parametric domain terrain model, and simultaneously generates a set of forward and inverse mapping relationships; The path filtering module receives a set of primary road alignments, a two-dimensional parameter domain terrain model, and a set of forward and inverse mapping relationships. It eliminates disconnected paths in the topological domain and filters out paths with excessive curvature or slope in the parameter domain, generating a path candidate strip that satisfies the dual-domain constraints. The reflection reconstruction module receives the path candidate strip and the set of forward and inverse mapping relationships, and transforms the path candidate strip to the original digital terrain grid coordinate system through the inverse Schwarz-Christoffel mapping to reconstruct the corresponding path shape. The parameter extraction module extracts the spatial projection length, segmented slope changes, and turning radius sequences based on the path alignment to construct a road geometry dataset.

[0018] In this embodiment, the modules are connected through the following method: A digital terrain grid is generated based on the elevation raster data, slope field data, and geomorphic feature data of the target wind farm area. Morse-Smale topological decomposition is performed on the digital terrain mesh to extract extreme points, saddle points and their connection relationships, and generate a set of topological skeleton lines. The local aspect consistency vector field is calculated based on the set of topological skeleton lines to generate a set of primary road alignments that express the natural travel direction; For a local terrain grid region covering the primary road alignment set, perform Schwarz–Christoffel conformal mapping to obtain a two-dimensional parametric domain terrain model, and simultaneously generate a set of forward and inverse mapping relationships; In the topology domain, primary lines that are not connected to the skeleton structure are removed. In the parameter domain, lines with abrupt curvature changes or slopes exceeding the threshold are identified and excluded. Based on the results of the two domains, path candidate zones are generated. By using the inverse mapping relationship, the path candidate band is mapped from the parameter domain back to the original digital terrain grid coordinate system, generating the path shape corresponding to the terrain; Based on the path alignment, the spatial projection length, segmented slope changes, and turning radius sequence are calculated to construct a geometric structure dataset for road design.

[0019] In this embodiment, generating a digital terrain mesh includes Read the elevation raster data covering the target wind farm area and index it according to two-dimensional latitude and longitude coordinates to construct the basic spatial framework of the grid cells; In the basic spatial framework, for the location of each grid cell, the corresponding slope field data is matched, the slope angle value and slope orientation information are extracted, and they are added to the current grid cell as slope attributes. Furthermore, in each grid cell, the geomorphic feature data label corresponding to its geographical location is read, and the label is written into the grid attribute structure as a geomorphic type field; In the grid structure, the elevation difference between adjacent grids is calculated according to the four-neighbor or eight-neighbor structure of each grid cell, and the local elevation change rate is calculated to characterize the local change trend of the cell. The elevation value, slope angle, slope direction, landform type, and local elevation change rate contained in each grid cell are written into the digital terrain data structure to construct a data grid model that is searchable, locatable, and has spatial differential features.

[0020] In this embodiment, generating the topology skeleton line set includes: Using a digital terrain grid as input, the elevation value of each grid cell is written into a scalar field array, and point-by-point gradient calculation is performed on the scalar field array, recording the gradient direction and gradient magnitude of each grid cell; For each grid cell in the scalar field array, compare its elevation relationship with that of its neighboring cells point by point. Grid cells with elevations higher than all neighboring cells are marked as maxima, those with elevations lower than all neighboring cells are marked as minima, and grid cells with abrupt changes in gradient direction and reverse changes in gradient magnitude are marked as saddle points, thus forming a list of key point indices. Input the list of key point indices into Morse-Smale topology decomposition, classify the key point numbers, and output a key point type table and a topology output table showing the connection relationships between key points. Read the key point number pairs recorded in the topology output table, perform line segment interpolation between each pair of key points according to the coordinate position of the number pairs in the digital terrain grid, and write the interpolated line segment sequence into the trajectory line list in spatial order. The list of trajectory lines is checked for breakpoints. Line segments that do not meet the continuity requirements are filled with coordinate points to ensure that each trajectory line maintains the same spatial resolution as the digital terrain grid. Duplicate segments are cleaned up and the sequence numbers are sorted for all trajectory lines. Establish a correspondence between the maximum point number, minimum point number, and saddle point number in the key point type table and the line segment sequence in the trajectory line list, and bind each trajectory line to its start and end key point number. The key point coordinates, key point numbers, key point types, and their corresponding trajectory line sequences together constitute a set of topological skeleton lines.

[0021] In this embodiment, the Morse-Smale topological decomposition includes: Using digital terrain grid as input, a unique index number is created for each grid cell, and its elevation value, gradient vector and adjacent cell index are written into a unified grid structure table. In the unified grid structure table, point-by-point gradient calculation is performed on all grid cells, gradient direction number is assigned to each grid cell, and the gradient direction number is written into the gradient direction index table. Based on the gradient direction index table, perform neighborhood elevation comparison on all grid cells, register grid cells that are higher than all neighboring cells as maximum points, register those that are lower than all neighboring cells as minimum points, and register grid cells whose neighborhood gradient direction shows a bidirectional inflection as saddle points, thus generating an initial set of key points. In the initial key point set, a direction sequence list containing the gradient direction numbers of its neighborhood is constructed for each key point and arranged in a fixed direction order. The length of the same gradient direction segment that appears consecutively in the direction sequence list is detected. When the length exceeds a preset threshold, the corresponding key point is removed from the initial key point set and the key point is registered as a transition unit for subsequent path segmentation. Write the filtered key points into a key point list, and record a unique key point number and type label for each key point; For each saddle point in the keypoint list, establish an ascending tracking buffer sequence and a descending tracking buffer sequence, and perform bidirectional gradient tracking as follows: During the ascent tracking process, starting from the saddle point unit, the process advances unit by unit according to the gradient direction number, and each advancing unit is written into the ascent tracking buffer sequence until the maximum value unit is reached or the preset maximum number of advancing steps is reached. During the descent tracking process, starting from the saddle point unit, the process proceeds in a direction symmetrical to the ascent tracking, writing each advancement unit into the descent tracking buffer sequence until the minimum point unit is reached or the preset maximum number of advancement steps is achieved. During the advancement process, when encountering a grid cell registered as a transition cell, local interpolation is performed based on the direction segment index interval recorded by the transition cell. The interpolated cell is then written into the current tracking buffer sequence, and the advancement operation continues. Each ascending tracking buffer sequence is registered as an ascending skeleton segment, and each descending tracking buffer sequence is registered as a descending skeleton segment. Each skeleton segment is then bound to its starting saddle point number and its ending key point number. Perform coordinate continuity checks on all skeleton segments. When an index jump is detected, perform linear interpolation to fill the segments. Clean up duplicate segments and sort the sequence numbers of the filled skeleton segments to ensure that all skeleton segments are arranged continuously in the grid space. Write the key point number, key point type label, skeleton segment index sequence, and skeleton segment connection relationship into the topology output table.

[0022] In this embodiment, the steps for generating the primary road alignment set include: Using the set of topological skeleton lines as input, the grid index sequence of each skeleton line is read, and the skeleton line coverage area is marked in the digital terrain grid. The skeleton line unit and its neighboring unit are written into the candidate region table. In the candidate region table, the slope aspect vector is calculated for each cell, where the slope aspect vector is the vector pointing from the current grid cell to the adjacent cell in the direction of maximum slope, and all slope aspect vectors are recorded in the slope aspect vector table; In the aspect vector table, the aspect vectors belonging to the skeleton line unit are assigned a direction number. The direction number is written into the skeleton direction table along with the grid index. The consistency of the direction number order of adjacent units in the skeleton direction table is checked. Units whose direction number changes more than a preset threshold are registered as direction transition units. Within the candidate region table, starting from each skeleton line unit, advance along its corresponding direction by numbering to construct a unit advancement sequence. Write the index of each unit in the advancement sequence into the direction advancement table. When advancing to a direction transition unit, read the slope vector of its adjacent unit, and select the unit with the smallest angle among the adjacent units as the next advancement unit based on the angle relationship between the slope vector and the current advancement direction. Write it into the direction advancement table and continue advancing. The propulsion sequences in all directions of the propulsion table are traversed. The grid index continuity is checked in each propulsion sequence. Linear interpolation is performed at the positions where index jumps occur. The interpolated cells are written into the propulsion sequence to keep the propulsion sequence continuous in the grid coordinate space. In each continuous advance sequence, record the starting skeleton line element index, the ending element index, and all the grid element indices contained therein. Define the sequence as a directional travel line element and write all directional travel line elements into the directional line element table. In the directional alignment table, the slope aspect vectors of all alignment units are checked for monotonicity. Alignment units whose slope aspect vector sequences satisfy the monotonicity condition are marked as primary road alignments, and alignment units that do not satisfy the monotonicity condition are deleted from the directional alignment table. All linear elements that pass the monotonicity check are combined according to their start and end element indices, element sequence, and direction number to form a primary road alignment set.

[0023] In this embodiment, generating the forward and reverse mapping relationship set includes: Using the primary road alignment set as input, the index range of the digital terrain grid covered by all alignment units is extracted, and a local terrain subgrid structure is established in the digital terrain grid. Each cell contains geographic coordinates. With elevation value ; In the local terrain subgrid, the center path containing the direction of maximum slope is determined as the reference baseline. Gaussian weighted smoothing is then applied to the elevation values ​​on both sides of the reference baseline. The smoothing weight function is defined as follows: ; in, The distance between the cell and the baseline. The smoothed radius coefficient is used to output the smoothed elevation matrix. Boundary point sets are generated using an elevation contour extraction algorithm. ; For the set of boundary points Execute the corner detection algorithm to calculate the exterior angle between three adjacent points. Register the angle abrupt change points as the boundary inflection point index sequence; Discrete sampling is performed within the boundary curve segment between adjacent inflection points. When the distance between adjacent sampling points is greater than a preset resolution threshold... When this happens, an auxiliary node is inserted at the midpoint to form a discrete boundary node set; For each boundary node, calculate the local slope direction vector in the original terrain mesh. And calculate the boundary normal direction vector. Define the node weighting function based on the angle relationship between the two vectors: , weight Write to the boundary weighting table, and according to Sort from largest to smallest; Select the top 75% of nodes from the sorting results to form the boundary skeleton node set. And connect them in spatial index order to form a conformal mapping boundary input sequence. ; Calculate the angle between adjacent nodes: ; And perform normalization to meet the requirements. Obtain the angle parameter table ; Establish an array of boundary nodes on the complex plane. ; Using the angle parameter table For input, construct the integral kernel function of the Schwarz–Christoffel conformal mapping: ; in, This is a scaling constant used to control the mapping scale; This is a translation constant used to control the plane position of the parameter domain; The starting point for integration is defined as the first node of the boundary sequence. Discretize the integral kernel function in a local grid subset. Inside, the coordinates of each unit Represented in complex form And perform numerical integration to solve: For the boundary elements, the Gauss-Legendre integral formula is used for discretization and calculation. ; For internal elements, the complex trapezoidal integral method is used to calculate the incremental integral between adjacent nodes; Calculated real part virtual part Construct a two-dimensional parameter domain coordinate matrix ; In the parameter domain coordinate matrix, the original grid index number of each node is recorded. Establish an index key-value mapping table And write all mapping results into the parameter domain node table; The parameter domain node table is divided into regions based on the mean terrain curvature, and the average curvature of each sub-region is calculated. The system automatically establishes parameter block boundaries and generates N parameter block sets. ; In each parameter block Internally, establish a node mapping index table. Use the original grid index number as the key Corresponding parameter domain coordinates As a value, write it into the forward mapping sub-table. ; After generating all forward mapping sub-tables, perform a reverse mapping construction operation on each parameter block: Read All key-value pairs in the data; Using parameter domain coordinates For key, the original grid index number For the value, create a reverse mapping sub-table. ; Perform two-dimensional Delaunay triangulation on the inverse key coordinates and establish an interpolation function. , satisfy any ; Overlap detection is performed on the boundaries of adjacent parameter blocks. If duplicate key values ​​are found, the mapping pair with the smallest interpolation error is retained, and the rest are deleted, forming a globally unique positive mapping table. Reverse mapping table ; In the global mapping table, numerical standardization is performed on the coordinates of the parameter domain, retaining four significant decimal places, and non-monotonic data is linearly resampled so that the forward and inverse mapping relationships satisfy the injective condition. In the reverse mapping table In the context of all interpolation functions The output is linearly normalized so that the coordinates after inverse mapping are consistent with the original grid space coordinates. Maintain a consistent standard; Randomly select several verification point pairs from a local grid subset. Calculate the angle retention error before and after mapping. And the side length ratio error: ; in, These are the included angles of adjacent sides before and after the mapping. For the corresponding side length; When satisfied , When this happens, mark the mapping set as a valid mapping; Index compression is performed on the effective mapping set, integrating the forward mapping table, reverse mapping table, and interpolation function set into a unified index structure. It also includes a storage parameter domain node coordinate matrix. Boundary vertex sequence and angle parameter table ; Unified index structure It is named a two-dimensional parametric domain terrain model.

[0024] In this embodiment, the generated path candidate band includes: Using the primary road alignment set as input, the original terrain grid index sequence contained in each alignment is read one by one, and the skeleton line number that has a common grid cell with the index sequence is retrieved in the topological skeleton line set; When there are no grid cells in the index sequence of a primary line shape that coincide with the set of topological skeleton lines, the line shape is marked as a topologically disconnected line shape and removed from the set of primary road line shapes, resulting in a set of line shapes that retain topological constraints. For each line shape in the set of line shapes that are retained for topological constraints, the line shape index sequence is mapped to the two-dimensional parameter domain coordinate system according to its grid index order, generating a parameter domain coordinate sequence that corresponds one-to-one with the line shape, and the parameter domain coordinate sequence is written into the parameter line shape table; In the parameter line table, for each parameter line, extract the coordinate points of three adjacent parameter domains in sequence, construct a list of polyline segment units segment by segment, and record the turning direction number of each polyline segment. For each parameter line shape, perform sequential traversal of the polyline segment cell list. When the turning direction number of a continuous polyline segment changes in the opposite direction, record the corresponding polyline segment position as a geometric anomalous node and mark the line shape containing the node as a direction change line shape. For each parameter line shape, calculate the ratio of the longitudinal change to the lateral distance between the coordinates of adjacent parameter domains according to their coordinate order, and write the ratio into the slope sequence. When more than a preset number of abnormal ratios appear consecutively in the slope sequence, the corresponding line shape is marked as an abnormal slope line shape. Parametric lines marked as abrupt direction changes or anomalous slopes are removed from the parametric lines table, forming a set of parametric domain constraint-preserved lines. Perform intersection matching between the topology constraint-preserved line set and the parameter domain constraint-preserved line set according to the line set number, retain only the line sets that exist in both sets, and generate a bi-domain consistent line set; For the linear shapes in the dual-domain consistent linear shape set, the spatial distance between adjacent linear shapes is expanded according to their index sequence in the original terrain grid, and multiple linear shapes located on both sides of the same topological skeleton line and with a spacing less than the preset width are merged into a continuous strip structure. Write the start and end indices of each strip structure, the set of line numbering contained therein, and the corresponding grid index range into the path candidate strip.

[0025] In this embodiment, generating the path alignment corresponding to the terrain includes: Using the path candidate band as input, read the two-dimensional parameter domain coordinate index sequence corresponding to each line shape in the path candidate band group by group, and query the inverse mapping function number index table of the parameter block to which it belongs based on its number; Within each parameter block, locate the two-dimensional coordinates of the current path point in the parameter domain, and call the pre-generated Delaunay triangulation structure within the parameter block to locate the point by indexing the triangular unit where the coordinates are located. For each path point in the parameter domain, perform centroid coordinate interpolation within its triangular cell to obtain the corresponding floating-point coordinate value in the original terrain grid; write the interpolation result into the inverse mapping result table in two-dimensional coordinate form, and retain the original parameter domain sequence number; Perform a sequence check on all path points in the inverse mapping result table. If the Euclidean distance between two consecutive points is greater than a set threshold, insert the midpoint coordinates into the line segment to supplement the subdivision density and generate a continuous and uniformly dense terrain path coordinate sequence. Traverse all coordinate points in the terrain path coordinate sequence, call the data reading interface of the original digital terrain grid, perform bilinear interpolation calculation for each point to obtain its elevation value, and combine the elevation value with the horizontal coordinate to construct a three-dimensional path point sequence. The three-dimensional path point sequence is renumbered according to its index order in the original parameter domain, and divided into several polyline segments. The start and end point numbers, original projected coordinates and elevation values ​​of each polyline segment are recorded. Perform continuity checks on each polyline segment. If there are spatial jumps (such as polyline angles exceeding a set threshold or sudden changes in spacing), insert auxiliary nodes in the middle of the polyline segment, update the start and end order of the polyline segment, and write them into the path correction list. The entire set of corrected polyline path segments is organized into a standard path line structure, which includes path number, original parameter domain number, start and end point number, corresponding three-dimensional coordinate sequence and its original terrain projection index information. All path linear structures are organized sequentially according to path number. The structure set, along with the path index table, the original parameter domain block number, and the correction flag, are written into the path linear set dataset.

[0026] In this embodiment, constructing the geometric structure dataset for road design includes: Using the path line set dataset as input, read the three-dimensional coordinate sequence in each path line structure one by one, and build a path node index table according to the path number and node order; For each path line, perform coordinate difference calculation on adjacent nodes to obtain the horizontal distance and elevation difference between adjacent nodes, and write the two results into the node distance table and the node elevation difference table. The cumulative summation of all node difference results in the node distance table is used to obtain a continuous projected length sequence of the path line shape, and the projected length sequence is associated with the corresponding path number and written into the length data table. Using the node elevation difference table and node distance table as input, slope calculation is performed on adjacent nodes to generate a segmented slope value sequence; when the slope sign of consecutive nodes changes, the location of the change point is registered as a slope turning point and its index number is marked in the path structure; For three consecutive nodes in the path, construct a local arc structure, calculate the arc radius based on the geometric relationship of the three points, record the arc radius sequence in the turning radius table, and mark the maximum and minimum radius points in each path; For each path, perform a synchronous traversal of the segmented slope sequence and turning radius sequence, detect the locations where slope changes and curvature changes occur simultaneously, mark such nodes as geometric control nodes, and write the node number and corresponding geometric parameters into the control node table; The length data table, segmented slope table, turning radius table and control node table are indexed and bound to the path number respectively to form a multi-dimensional geometric parameter matrix structure; In the multidimensional geometric parameter matrix, linear interpolation smoothing is performed on all parameters of the same path to maintain a continuous and solvable relationship in space for the path length, slope and turning radius; The smoothed multidimensional geometric parameter matrix is ​​summarized sequentially according to the path number to generate a road geometry dataset containing a sequence of path projection lengths, a sequence of segmented slopes, a sequence of turning radii, and an index of geometric control nodes. Write the road geometry dataset into an output file structure and use it as input data for the road optimization design module.

[0027] Example: In a typical high-altitude wind farm road design and construction scenario, the terrain is extremely complex, with large elevation differences, dramatic slope changes, and irregular landforms. Traditional design methods based on manual route selection struggle to achieve coordinated control of slope and curves within a reasonable timeframe. To address the shortcomings of traditional methods in terms of path continuity, slope control, and curvature smoothness, this invention proposes a high-altitude wind farm road slope and curve optimization design system and method for practical application.

[0028] During implementation, the system first receives raw elevation raster data, slope field data, and geomorphic feature data from a Digital Elevation Model (DEM). A digital terrain mesh with differential properties is then established using the terrain modeling module. This module utilizes a set of high-precision interpolation algorithms to convert the raw elevation values ​​into a differentiable continuous surface, enabling the system to continuously analyze the local undulation trends, slope variations, and geomorphic features of the terrain. The terrain model output at this stage includes three-dimensional attributes such as node coordinates, elevation gradient, and local curvature, providing complete spatial information for subsequent topology extraction.

[0029] In the topology deconstruction module, the system performs structural analysis on the digital terrain mesh using the Morse-Smale topology decomposition method. The algorithm first calculates the gradient direction and magnitude of each grid cell, then identifies and marks maxima, minima, and saddle points. The module utilizes a gradient tracing algorithm to establish connections between key points, forming a set of topological skeleton lines. This set reflects the global structural characteristics of the terrain, including ridges, valleys, and slope variation trends. Compared to traditional route selection methods based on discrete elevation difference analysis, this topological skeleton structure can mathematically accurately represent the directionality of the terrain energy field, enabling road route selection to possess physical consistency and geometric interpretability.

[0030] After generating the topological skeleton, the system enters the orientation field generation module. This module reads the skeleton line index sequence, calculates the aspect vector for each grid cell, and adjusts the aspect consistency using a local slope gradient correction algorithm. The system constructs an orientation advancement table for each skeleton line, generating an aspect consistency vector field along the direction of maximum gradient, thus forming a primary set of road alignments for natural traffic directions. This set can automatically avoid terrain extrema areas, ensuring path continuity and conforming to the natural orientation of the terrain.

[0031] The mapping module takes a primary set of linear shapes as input and performs Schwarz–Christoffel conformal mapping on a local terrain mesh containing all linear shapes. The module utilizes a complex plane mapping kernel function to transform complex 3D terrain regions into a 2D parameter domain and simultaneously generates forward and inverse mapping tables. To ensure bidirectional mapping accuracy, the system establishes a node mapping index table within each parameter block and performs Delaunay triangulation, maintaining the local geometric continuity of the mapped region through interpolation functions. Verification shows that the mapping angle error is controlled within 2°, and the side length ratio error is less than 3%.

[0032] In the path selection module, the system performs path constraint operations simultaneously in both the topology domain and the parameter domain. The topology domain is used to eliminate paths that are not connected to the terrain skeleton, while the parameter domain excludes paths with abrupt curvature changes and abnormal slopes. The system retains a set of topologically consistent and geometrically reasonable paths through bi-domain intersection matching. The path selection algorithm calculates the local rate of curvature change and rate of slope change on each path, automatically deleting segments exceeding thresholds, resulting in a final path with continuous slope transitions and smooth turning structures.

[0033] The path reconstruction module receives the path set after dual-domain filtering and uses the inverse Schwarz–Christoffel mapping to remap the two-dimensional parameter domain paths back to the original terrain grid coordinate system. To ensure spatial accuracy, the system employs bilinear interpolation and midpoint supplementation mechanisms during the reconstruction process, maintaining uniform path density and geometric continuity in space. The coordinates of each path point are calculated using elevation interpolation to generate a complete three-dimensional path point sequence. The module records the path number, start and end nodes, elevation sequence, and projected coordinates in the output structure, providing input for the geometric parameter extraction module.

[0034] In the parameter extraction module, the system calculates the spatial projection length, segmented slope, and turning radius based on the path point sequence, generating a road geometry dataset. The curvature radius of each turning segment is obtained using a three-point circular arc fitting algorithm, and then the minimum, average, and maximum radii are calculated. The system interpolates and smooths the segmented slope sequence of the path, ensuring that the curvature and slope are spatially continuous and solvable. The final generated multidimensional geometric parameter matrix includes the projection length sequence, slope sequence, curvature sequence, and control node index, realizing a digital representation of the road structure.

[0035] In application testing, a mountainous area of ​​approximately 40 square kilometers was selected for system verification. Traditional methods require manual path determination, with an average design cycle of 28 days, while this system can output a complete road plan in just 5 days with fully automated operation. The maximum gradient of the path generated by the system is controlled at 11.5%, and the average gradient is 7.9%, which is significantly lower than the 15.2% of the traditional method. The average turning radius of the path is increased to 35.4 meters, and the proportion of sections with continuous curvature changes is increased to 92%.

[0036] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A road slope and curve optimization design system for high-altitude wind farms, characterized in that, include: The terrain modeling module takes regional elevation raster data, slope field data, and geomorphic feature data as input and generates a digital terrain grid. The topology deconstruction module receives the digital terrain mesh, performs Morse-Smale topology decomposition, extracts extreme points, saddle points and their connection structures, and constructs a set of topology skeleton lines. The orientation field generation module constructs a slope-aspect consistent vector field based on the topological skeleton line set, generating a primary road alignment set; The mapping construction module selects a local grid region containing the primary road alignment set, performs Schwarz–Christoffel conformal mapping, generates a two-dimensional parametric domain terrain model, and simultaneously generates a set of forward and inverse mapping relationships; The path filtering module receives a set of primary road alignments, a two-dimensional parameter domain terrain model, and a set of forward and inverse mapping relationships. It eliminates disconnected paths in the topological domain and filters out paths with excessive curvature or slope in the parameter domain, generating a path candidate strip that satisfies the dual-domain constraints. The reflection reconstruction module receives the path candidate strip and the set of forward and inverse mapping relationships, and transforms the path candidate strip to the original digital terrain grid coordinate system through the inverse Schwarz-Christoffel mapping to reconstruct the corresponding path shape. The parameter extraction module extracts the spatial projection length, segmented slope changes, and turning radius sequences based on the path alignment to construct a road geometry dataset.

2. A method for optimizing the design of road slope and curves in high-altitude wind farms, characterized in that, The modules are connected in the following way: A digital terrain grid is generated based on the elevation raster data, slope field data, and geomorphic feature data of the target wind farm area. Morse-Smale topological decomposition is performed on the digital terrain mesh to extract extreme points, saddle points and their connection relationships, and generate a set of topological skeleton lines. The local aspect consistency vector field is calculated based on the set of topological skeleton lines to generate a set of primary road alignments that express the natural travel direction; For a local terrain grid region covering the primary road alignment set, perform Schwarz–Christoffel conformal mapping to obtain a two-dimensional parametric domain terrain model, and simultaneously generate a set of forward and inverse mapping relationships; In the topology domain, primary lines that are not connected to the skeleton structure are removed. In the parameter domain, lines with abrupt curvature changes or slopes exceeding the threshold are identified and excluded. Based on the results of the two domains, path candidate zones are generated. By using the inverse mapping relationship, the path candidate band is mapped from the parameter domain back to the original digital terrain grid coordinate system, generating the path shape corresponding to the terrain; Based on the path alignment, the spatial projection length, segmented slope changes, and turning radius sequence are calculated to construct a geometric structure dataset for road design.

3. The method for optimizing the design of road slope and curves in a high-altitude wind farm according to claim 2, characterized in that, The generation of the topology skeleton line set includes: Using a digital terrain grid as input, the elevation value of each grid cell is written into a scalar field array, and point-by-point gradient calculation is performed on the scalar field array, recording the gradient direction and gradient magnitude of each grid cell; For each grid cell in the scalar field array, compare its elevation relationship with that of its neighboring cells point by point. Grid cells with elevations higher than all neighboring cells are marked as maxima, those with elevations lower than all neighboring cells are marked as minima, and grid cells with abrupt changes in gradient direction and reverse changes in gradient magnitude are marked as saddle points, thus forming a list of key point indices. Input the list of key point indices into Morse-Smale topology decomposition, classify the key point numbers, and output a key point type table and a topology output table showing the connection relationships between key points. Read the key point number pairs recorded in the topology output table, perform line segment interpolation between each pair of key points according to the coordinate position of the number pairs in the digital terrain grid, and write the interpolated line segment sequence into the trajectory line list in spatial order. The list of trajectory lines is checked for breakpoints. Line segments that do not meet the continuity requirements are filled with coordinate points to ensure that each trajectory line maintains the same spatial resolution as the digital terrain grid. Duplicate segments are cleaned up and the sequence numbers are sorted for all trajectory lines. Establish a correspondence between the maximum point number, minimum point number, and saddle point number in the key point type table and the line segment sequence in the trajectory line list, and bind each trajectory line to its start and end key point number. The key point coordinates, key point numbers, key point types, and their corresponding trajectory line sequences together constitute a set of topological skeleton lines.

4. The method for optimizing the design of road slope and curves in a high-altitude wind farm according to claim 2, characterized in that, Generating the primary road alignment set includes: Using the set of topological skeleton lines as input, the grid index sequence of each skeleton line is read, and the skeleton line coverage area is marked in the digital terrain grid. The skeleton line unit and its neighboring unit are written into the candidate region table. In the candidate region table, the slope aspect vector is calculated for each cell, where the slope aspect vector is the vector pointing from the current grid cell to the adjacent cell in the direction of maximum slope, and all slope aspect vectors are recorded in the slope aspect vector table; In the aspect vector table, the aspect vectors belonging to the skeleton line unit are assigned a direction number. The direction number is written into the skeleton direction table along with the grid index. The consistency of the direction number order of adjacent units in the skeleton direction table is checked. Units whose direction number changes more than a preset threshold are registered as direction transition units. Within the candidate region table, starting from each skeleton line unit, advance along its corresponding direction by numbering to construct a unit advancement sequence. Write the index of each unit in the advancement sequence into the direction advancement table. When advancing to a direction transition unit, read the slope vector of its adjacent unit, and select the unit with the smallest angle among the adjacent units as the next advancement unit based on the angle relationship between the slope vector and the current advancement direction. Write it into the direction advancement table and continue advancing. The propulsion sequences in all directions of the propulsion table are traversed. The grid index continuity is checked in each propulsion sequence. Linear interpolation is performed at the positions where index jumps occur. The interpolated cells are written into the propulsion sequence to keep the propulsion sequence continuous in the grid coordinate space. In each continuous advance sequence, record the starting skeleton line element index, the ending element index, and all the grid element indices contained therein. Define the sequence as a directional travel line element and write all directional travel line elements into the directional line element table. In the directional alignment table, the slope aspect vectors of all alignment units are checked for monotonicity. Alignment units whose slope aspect vector sequences satisfy the monotonicity condition are marked as primary road alignments, and alignment units that do not satisfy the monotonicity condition are deleted from the directional alignment table. All linear elements that pass the monotonicity check are combined according to their start and end element indices, element sequence, and direction number to form a primary road alignment set.

5. The method for optimizing the design of road slope and curves in a high-altitude wind farm according to claim 2, characterized in that, Generating sets of forward and reverse mapping relationships includes: Using the set of primary road alignments as input, the index range of the digital terrain grid containing all primary alignments is extracted. The corresponding local terrain region is located in the digital terrain grid, and the elevation values ​​of all grid cells in the region are read to form a local terrain data matrix. In the local terrain data matrix, a linear region containing the direction of maximum slope is selected as the reference baseline. Grid-weighted smoothing is performed on the elevation data on both sides of the reference baseline to generate a set of boundary isopleths, and the set of isopleths is used as the input of local boundary data. Perform corner detection on local boundary data to determine the boundary inflection point index sequence. and the included angle between adjacent boundary segments Piecewise discretization is performed to decompose the continuous curve boundary into a finite piecewise line segment structure; Insert auxiliary nodes within each boundary polyline segment, ensuring that the distance between adjacent nodes within each segment is less than a preset boundary resolution threshold. And calculate the local normal direction vector of each node. ; Based on the angle between the normal direction of the boundary node and the local slope direction Establish node weighting functions , weight Write to the boundary weighting table; In the boundary weighted table, all nodes are sorted by weight. Sort the nodes from largest to smallest, select a fixed proportion of the node set to form the boundary skeleton node set, and connect them in the order of this node set to form the input boundary vertex sequence of the conformal mapping. The angle between adjacent line segments of the boundary vertex sequence Write the angle parameter table and perform normalization; Using the boundary vertex sequence and angle parameter table as input, construct a complex plane boundary vertex array. ,in Represented by complex coordinates; Define the integral kernel function for the Schwarz–Christoffel conformal mapping; Establish a mapping integral solution grid, and assign coordinates of each node within a subset of the local terrain grid. Mapped to complex plane coordinates And perform integration on each node: Gaussian quadrature is applied to the boundary nodes; The trapezoidal integral method is used for internal nodes; The real and imaginary parts of the integral result are represented as two-dimensional parameter domain coordinates. ; All nodes in the parameter field Coordinates and their corresponding original grid index numbers Write the parameters to the parameter field node table and index them in spatial order; Divide the parameter domain node table into Each parameter block contains a fixed number of nodes; Within each parameter block, a node index mapping table is established, using the original mesh index number as the key and the parameter domain coordinates as the key. As the value, write it to the forward mapping sub-table, and mark the sub-table for each parameter block as... ; After generating all forward mapping sub-tables, for each parameter block Perform reverse matching and construction operations: Read sub-table All key-value pairs in the database; Using parameter domain coordinates Using the key and the original index number as the value, a reverse mapping sub-table is generated. ; Perform two-dimensional Delaunay triangulation on the inverse coordinate keys within each parameter block, and establish a coordinate interpolation function based on the triangulation results. ; Perform overlapping boundary detection on all parameter blocks. When duplicate mapping points are detected at the boundary of a parameter block, retain the key-value pair with the smallest mapping error and delete the rest of the mapping records to form a unified forward mapping table and reverse mapping table. In the global forward mapping table, the mapped coordinate data is standardized to floating-point values. In the global inverse mapping table, for all inverse interpolation functions The output is linearly normalized once to ensure that the mapped coordinate values ​​are at the same scale as the original grid space. After completion, the forward mapping table, reverse mapping table and interpolation function set are merged to form a mapping relationship set, and its parameter domain boundary index, number of nodes and coordinate distribution range are recorded; Within a subset of the local terrain grid, N pairs of mapping verification points are randomly selected. For each verification point, the angle preservation error after mapping is calculated. Error compared to side length When satisfied and When this happens, the mapping set is marked as valid. and To set a threshold; Index compression is performed on the effective mapping relationship set, merging the forward and reverse mapping tables into a unified index structure, and recording the parameter domain node coordinates, boundary vertex sequences and angle parameter tables in the same data block; The above unified index structure is named the two-dimensional parameter domain terrain model.

6. The method for optimizing the design of road slope and curves in a high-altitude wind farm according to claim 2, characterized in that, The generated path candidate bands include: Using the primary road alignment set as input, the original terrain grid index sequence contained in each alignment is read one by one, and the skeleton line number that has a common grid cell with the index sequence is retrieved in the topological skeleton line set; When there are no grid cells in the index sequence of a primary line shape that coincide with the set of topological skeleton lines, the line shape is marked as a topologically disconnected line shape and removed from the set of primary road line shapes, resulting in a set of line shapes that retain topological constraints. For each line shape in the set of line shapes that are retained for topological constraints, the line shape index sequence is mapped to the two-dimensional parameter domain coordinate system according to its grid index order, generating a parameter domain coordinate sequence that corresponds one-to-one with the line shape, and the parameter domain coordinate sequence is written into the parameter line shape table; In the parameter line table, for each parameter line, extract the coordinate points of three adjacent parameter domains in sequence, construct a list of polyline segment units segment by segment, and record the turning direction number of each polyline segment. For each parametric line shape, perform sequential traversal of the polyline segment cell list. When the turning direction number of a continuous polyline segment changes in the opposite direction, record the corresponding polyline segment position as a geometric anomalous node and mark the line shape containing the node as a direction change line shape. For each parameter line shape, calculate the ratio of the longitudinal change to the lateral distance between the coordinates of adjacent parameter domains according to their coordinate order, and write the ratio into the slope sequence. When more than a preset number of abnormal ratios appear consecutively in the slope sequence, the corresponding line shape is marked as an abnormal slope line shape. Parametric lines marked as abrupt direction changes or anomalous slopes are removed from the parametric lines table, forming a set of parametric domain constraint-preserved lines. Perform intersection matching between the topology constraint-preserved line set and the parameter domain constraint-preserved line set according to the line set number, retain only the line sets that exist in both sets, and generate a bi-domain consistent line set; For the linear shapes in the dual-domain consistent linear shape set, the spatial distance between adjacent linear shapes is expanded according to their index sequence in the original terrain grid, and multiple linear shapes located on both sides of the same topological skeleton line and with a spacing less than the preset width are merged into a continuous strip structure. Write the start and end indices of each strip structure, the set of line numbering contained therein, and the corresponding grid index range into the path candidate strip.

7. The method for optimizing the design of road slope and curves in a high-altitude wind farm according to claim 2, characterized in that, Generating path alignments corresponding to the terrain includes: based on the coordinate index sequence of each alignment in the parameter domain within the candidate path zone, using the inverse mapping relationship of the corresponding parameter blocks, mapping the path points in the parameter domain back to the original digital terrain grid coordinate system through centroid interpolation under Delaunay triangulation, performing spacing supplementation and elevation interpolation on the mapped path points, constructing a continuous three-dimensional path point sequence according to the original parameter domain order, and organizing the set of polyline segments after continuity correction into a path alignment set dataset containing three-dimensional coordinate information according to the path number.

8. The method for optimizing the design of road slope and curves in a high-altitude wind farm according to claim 2, characterized in that, The construction of a geometric structure dataset for road design includes: based on the three-dimensional coordinate sequence of each path in the path alignment set, calculating and recording the cumulative spatial projection length of each path, the segmented slope change value between adjacent nodes, and the turning radius sequence determined by three consecutive nodes in sequence, and storing the spatial projection length, segmented slope change and turning radius sequence in association according to the path number to form a geometric structure dataset for road design.

Images