Intelligent generation method of urban three-dimensional road network for complex mountain terrain

Abstract

The invention discloses an intelligent generation method of an urban three-dimensional road network for a complex mountain terrain. The method comprises the steps of data acquisition and processing of a design site, node identification of a primary road network, generation of the primary road network, generation of a secondary road network, road network smoothing and included angle optimization, and road network output and interactive optimization. According to the invention, the method carries out the path search and optimization through the unified processing of the current terrain and environment data of a site in combination with an improved A * algorithm, intelligently generates a primary and secondary road network, and carries out the self-adaptive smoothing processing. The finally output three-dimensional road network has higher adaptability and rationality, interactive optimization display is more visual and stereoscopic, and effective support is provided for urban road planning of complex terrains.

Description

Technical Field

[0001] The present invention belongs to the technical field of urban and rural intelligent planning, and in particular relates to an intelligent generation method of a three-dimensional urban road network for complex mountainous terrain. Background Art

[0002] Urban road network layout is an important part of urban planning and an important means of controlling urban spatial structure, which profoundly affects the city's operating efficiency and future development. Road network generation technology mainly has two application scenarios: new city construction road planning and old city expansion road planning. In actual application scenarios, it is often faced with problems such as complex terrain, dense existing reserves, and high construction costs. Traditional road network generation methods are difficult to balance the relationship between complex existing conditions and planning intentions, and are difficult to adapt to the design requirements of road networks in different scenarios. It is also impossible to dynamically and intuitively interactively adjust and improve the generation scheme.

[0003] Therefore, it is urgent to solve the above problems. Summary of the invention

[0004] Purpose of the invention: The purpose of the present invention is to provide an intelligent method for generating urban three-dimensional road networks for complex mountainous terrains, which solves the problems that existing urban road network generation methods are difficult to adapt to complex multiple scenarios, difficult to balance the relationship between complex current conditions and multiple factors such as planning intentions, and unable to dynamically and interactively adjust and improve the generation scheme.

[0005] Technical solution: To achieve the above objectives, the present invention discloses an intelligent generation method of a three-dimensional urban road network for complex mountainous terrain, comprising the following steps:

[0006] (1) Collect the current terrain data and environmental data of the design site, collect the relevant upper-level planning data of the design site, import them into the geographic information platform, process them into raster data or vector data in a unified format, and generate the design site database;

[0007] (2) Based on the design site database, the development suitability of each area of ​​the design site is calculated, and after threshold screening, the center point of each area is identified as the main road node in the high suitability area;

[0008] (3) Generate a three-dimensional grid of the design site, use the improved A* algorithm to search for paths, connect the main road nodes and extend them to the edge of the design site to form several main roads, delete unreasonable paths, calculate the distance between the main roads, and make supplementary adjustments to the road network;

[0009] (4) Based on the generated trunk road network, secondary trunk road nodes are selected on the trunk road according to the road spacing specification, and the improved A* algorithm is used to perform path search to generate the secondary trunk road network;

[0010] (5) Use an adaptive smoothing algorithm to process the generated primary and secondary road networks and optimize the road angles;

[0011] (6) Output the generated road network and perform three-dimensional visualization interactive optimization.

[0012] Optionally, step (1) specifically includes the following steps:

[0013] (1.1) Collect the current terrain data of the design site and obtain DEM raster data through the geospatial data cloud or the geological survey bureau in the city;

[0014] (1.2) Collect the current environmental data of the design site, including vector environmental data of rivers and current buildings, atmospheric environmental data and raster data of land cover types;

[0015] (1.3) Use drones equipped with high-resolution cameras and LiDAR to take aerial images of the design site, and identify the current terrain data of the design site and the current environmental data of the design site (excluding atmospheric environmental data) from the aerial images to assist in improving the current terrain data and environmental data of the design site;

[0016] (1.4) Obtain the superior planning data of the design site from the local urban planning department. The superior planning data includes the city master plan, regional development plan, transportation plan and superior planning cultivated land indicators;

[0017] (1.5) Import the collected terrain data, environmental data and upper-level planning data into a high-performance workstation, use the geographic information platform, select the coordinate system based on the site coordinates, add geographic registration control points, and spatially locate the upper-level planning data; convert the planning map of the upper-level planning data from raster data to vector elements in the geographic information platform, correct the errors in the conversion process, identify the scope of various functional areas in the planning map, assign values ​​to different functional areas, and store them as raster data;

[0018] (1.6) The data is integrated to generate the design site database, which is stored in the RAID 10 storage array and managed using a database manager.

[0019] Optionally, in step (1.2), vector environmental data of rivers and existing buildings are obtained through an open source data platform; atmospheric environmental data are obtained through the China Meteorological Data Network; and raster data of surface cover types are obtained through the National Geographic Information Resource Directory Service System.

[0020] Optionally, step (2) specifically includes the following steps:

[0021] (2.1) Based on the design site database, the design site is analyzed using the geographic information platform to obtain the evaluation results of the normalized slope index, regional climate comfort index, land transformation index, and functional area transformation time series index;

[0022] (2.2) According to the analytic hierarchy process (AHP), a judgment matrix is ​​constructed to obtain the weight of each factor; the slope index, regional climate comfort index, land transformation index, and functional area transformation time series index are superimposed and calculated according to the weights calculated by the judgment matrix to obtain the basic map of the development suitability of the design site and store it in raster data format;

[0023] (2.3) Establish a grid, convert the development suitability base map into point data, use the spatial connection tool to extract the value of the point data into the corresponding grid, use the natural segmentation method to classify the extracted data into 5 levels, from 1 to 5 corresponding to low to high development suitability, and store the classification results in a new field of "development suitability"; use the attribute selection tool in the attribute table to filter out the cells with a value of 5 in the development suitability field, that is, cells with high development suitability, and take the geometric center of each high suitability area as the main road node;

[0024] (2.4) Check whether the location of each node has steep slopes or terrain that is not suitable for development. If so, move the node to the nearest flat area.

[0025] Optionally, step (2.1) specifically includes the following steps:

[0026] The slope was calculated using DEM raster data, and the slope was evaluated. It was divided into five levels: high, relatively high, medium, relatively low, and low, with scores of 1, 2, 3, 4, and 5, according to ≤2°, 2-6°, 6-15°, 15-25°, and >25°. The high level corresponds to flat land, the relatively high level corresponds to flat slope land, the medium level corresponds to gentle slope land, the relatively low level corresponds to gentle steep slope land, and the relatively low level corresponds to steep slope land. A slope classification map was generated.

[0027] The atmospheric environment data is used to calculate the average monthly temperature and the average monthly relative humidity, calculate the temperature and humidity index, and evaluate the regional climate comfort level. The regional climate comfort level is divided into 7 levels and scored according to the level. The temperature and humidity index of 60-65 corresponds to the comfort level of 7, with 7 points; the temperature and humidity index of 56-60 or 65-70 corresponds to the comfort level of 6, with 6 points; the temperature and humidity index of 50-56 or 70-75 corresponds to the comfort level of 5, with 5 points; the temperature and humidity index of 45-50 or 75-80 corresponds to the comfort level of 4, with 4 points; the temperature and humidity index of 40-45 or 80-85 corresponds to the comfort level of 3, with 3 points; the temperature and humidity index of 32-40 or 85-90 corresponds to the comfort level of 2, with 2 points; the temperature and humidity index <32 or >90 corresponds to the comfort level of 1, with 1 point; the higher the comfort level, the higher the comfort;

[0028] Using land cover type data and based on the upper-level planned cultivated land indicators, the degree of land transformation is evaluated, and each type of land is divided into three levels: transformable, moderately transformable, and non-transformable, and each is assigned a score of 5, 3, and 1.

[0029] Using the city master plan, each functional area is divided into three levels: pending development, near-term development and key development, and is scored 1, 2 and 3 points;

[0030] The four evaluation scores were normalized to obtain the slope index, regional climate comfort index, land transformability index, and functional area transformation time series index.

[0031] Optionally, the order of importance of the judgment matrix in step (2.2) is: slope index > land transformability index > functional area transformation time series index > regional climate comfort index; calculate the sum of each column of the judgment matrix; divide each element by its column sum to obtain a normalized judgment matrix; then calculate the average value of each row to obtain the weight of each factor;

[0032] The superposition calculation formula for development suitability is: k1×w1+k2×w2+k3×w3+k4×w4, where k1 is the slope index, w1 is the weight of the slope index, k2 is the regional climate comfort index, w2 is the weight of the regional climate comfort index; k3 is the land transformability index, w3 is the weight of the land transformability index; k4 is the functional area transformation timing index, and w4 is the weight of the functional area transformation timing index.

[0033] Optionally, step (3) specifically includes the following steps:

[0034] (3.1) Identify the center point of each grid of the design site DEM grid data containing height coordinates, connect all grid center points to generate a three-dimensional grid, use a high-speed storage device NVMe SSD to write the three-dimensional grid, and use it to generate a road network storage;

[0035] (3.2) Use the improved A* algorithm to search for paths, define the main road nodes as the starting point and end point of the path search, and the formula of the function f(n) of the improved A* algorithm is:

[0036] f(n)=g(n)+w·h(n)+α·S(n)+β·C(n)

[0037] g(n) is the actual cost from the starting point to the current node n, that is, the path length;

[0038] h(n) is the Euclidean distance heuristic function, which estimates the cost from node n to the target node;

[0039] w is a dynamic weight, which is used to adjust the influence of h(n) and changes dynamically according to the progress of the search;

[0040] S(n) is the terrain cost function, which reflects the slope or terrain undulation of the area where node n is located;

[0041] α is the weight of the terrain cost, which determines the impact of terrain on path selection;

[0042] C(n) is the cost function of the road design rule, which takes into account the factors of road spacing and road curvature;

[0043] β is the weight of the road design rule, which determines the impact of the road design rule on path selection;

[0044] in,

[0045] g(n)=g(p)+d(p,n)

[0046] g(p) is the cost of the previous node p;

[0047] d(p,n) is the path length from node p to node n;

[0048]

[0049] Δh(n) is the height difference from the previous node p to the current node n, i.e. the slope;

[0050] d(p,n) is the horizontal distance from the previous node p to the current node n;

[0051]

[0052] (x n ,yn ,z n ) is the three-dimensional coordinate of the current node n;

[0053] (x goal ,y goal ,z goal ) is the three-dimensional coordinate of the target node;

[0054]

[0055] d road (n) is the actual distance between the road at the current node n and the nearest adjacent road;

[0056] d target The ideal road spacing is predefined, which is 800-1000 meters according to urban design specifications;

[0057] θ(n) is the turning angle of the current path at node n, expressed in degrees;

[0058] θmax is the maximum allowed road turning angle, which is determined to be 60 degrees according to the design standard;

[0059] (3.3) Use the improved A* algorithm in step (3.2) to perform path search, define the trunk road node as the starting point of the path search, design the site 3D grid boundary point as the end point of the path search, generate a path with the minimum f(n) from each trunk road node n to the grid boundary, and preliminarily form a trunk road network;

[0060] (3.4) Check the node degree of the main road nodes after the road network is generated, identify all the paths connected to the nodes whose degree exceeds the expected value of 4, and prioritize them. The path connecting the grid boundary has the highest priority, and the remaining paths are ranked based on the cost function. The cost function comprehensively considers the slope and distance, and deletes the path with the lowest priority until the node degree is less than or equal to 4;

[0061] The cost function formula is:

[0062]

[0063] w1 and w2 are weight coefficients, which control the influence of slope cost and distance cost respectively;

[0064] Δh r is the height difference between the starting point and the end point of path r;

[0065] d r is the length of path r;

[0066] (3.5) Calculate the spacing between the generated trunk road networks, and mark all areas with a spacing greater than 1200 meters according to the trunk road spacing standard; in each marked area, calculate its geometric center point as a new node, and use the improved A* algorithm in step (3.2) to search for a path from this node to the existing trunk road node with the longest distance among all trunk roads at the edge of the area to obtain a connecting path; extend the trunk roads from both ends of the connecting path to the three-dimensional grid boundary of the design site to form several trunk roads.

[0067] Optionally, step (4) specifically includes the following steps:

[0068] (4.1) Select secondary trunk road nodes on the generated trunk road network according to the spacing rule. The spacing rule is greater than 350 meters and less than 500 meters, and the closer to the three-dimensional grid boundary, the larger the spacing;

[0069] (4.2) For each area divided by the trunk road network, use the improved A* algorithm in step (3.2) to search for paths, adjust the function formula of the improved A* algorithm, increase the weight of the road design rule cost, that is, increase the β value in the formula, define the secondary trunk road nodes on the boundary of the area as the starting point and end point, and the nodes on the same trunk road path cannot be used as a pair of starting points and end points;

[0070] (4.3) After the path is generated, the node degree of the secondary trunk road nodes is checked, and the nodes whose degree exceeds the expected value of 1 are identified. The main road network paths and paths outside the area are not calculated, and the redundant paths are deleted so that the degree of all nodes is less than or equal to 1, and the secondary trunk road network is obtained.

[0071] Optionally, step (5) specifically includes the following steps:

[0072] (5.1) Divide the road network into different areas. According to the terrain characteristics, the road network is divided into flat areas with a slope of 0°-5°, hilly areas with a slope of 5°-15°, and mountainous areas with a slope of more than 15°;

[0073] (5.2) Dynamically select and adjust the smoothing algorithm and window according to different regions. The window represents the number of adjacent data points selected within a given range. For flat areas, the linear interpolation method is selected, and the parameters use a larger smoothing window; for hilly areas, the Bezier curve is selected, and the parameters use a medium-sized smoothing window; for mountainous areas, the Gaussian filter is used, and the parameters use a smaller smoothing window;

[0074] (5.3) Convert the smoothed road network into a multidimensional vector data set, use Python to calculate the direction vector, and calculate the angle based on the direction vector;

[0075] (5.4) Identify the intersection node A with an angle less than 70 degrees, and divide the two direction vectors involved in the angle into the main section and the secondary section; take this node as the starting point, and generate an alternative node B on the main section with a step length of cot(θ)*10m; generate a new section perpendicular to the main section from the alternative node B, and calculate the intersection C of the new section and the secondary section; recalculate the local direction vector and the angle, and repeat the calculation until the road angle is greater than 70 degrees;

[0076] (5.5) Write the generated road network into a high-speed storage device.

[0077] Optionally, step (6) specifically includes the following steps:

[0078] (6.1) Input the generated road network data into the virtual reality application software on the development workstation equipped with high-performance CPU and GPU for processing and rendering;

[0079] (6.2) Connect VR headsets and supporting controllers, integrate and test different VR devices, and create VR scenes; connect AR devices, and place and track virtual objects through mobile devices or AR glasses; perform 3D visualization and interactive optimization of the generated road network on high-resolution displays of 4K and above, and use VR and AR tracking devices to adjust and optimize the scene.

[0080] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:

[0081] (1) The present invention collects relevant design site vector data and inputs it into the geographic information platform for format unification processing, constructs a design site digital sand table, and forms a design site condition data set, thereby ensuring the standardization and accuracy of the data and improving the accuracy and efficiency of urban road network generation;

[0082] (2) The present invention analyzes multiple factors such as terrain slope, climate comfort, and surface cover type, and comprehensively calculates and standardizes the "development suitability" of the region. It effectively selects high-suitability areas as trunk road nodes, so that the subsequent generation of trunk roads and secondary roads is based on more scientific and reasonable point selection, thereby improving the planning quality of the road network. This technology ensures that the generated road network not only conforms to the characteristics of the regional natural environment, but also meets the actual needs of regional development, providing a solid data foundation and spatial basis for subsequent road network planning, and improving the overall planning efficiency and accuracy;

[0083] (3) The present invention takes into account multiple factors such as path length, terrain complexity, and road design rules through the improved A* algorithm, making the planning of trunk roads more accurate; the introduced weight parameters can be dynamically adjusted according to actual needs, so that the generated trunk roads are more in line with the terrain characteristics, avoiding the problem of insufficient processing of complex terrain in traditional path algorithms; in addition, the application of node priority sorting and degree value detection reduces the calculation of redundant paths and improves the efficiency of the trunk road generation process, so that the algorithm can maintain high calculation efficiency and path rationality when running in large-scale three-dimensional grid data;

[0084] (4) The present invention adopts a partition optimization search strategy based on the trunk road to ensure that the planning of secondary trunk roads follows the overall structure of the trunk road, and through reasonable weight parameter setting, the secondary trunk roads can adapt to terrain changes and specific road requirements in different areas; through the degree value detection and priority sorting of the secondary trunk road nodes, the redundant paths are effectively removed and the structure of the secondary trunk road network is simplified, so that the generated secondary trunk roads can both reasonably connect the trunk road network and maintain a high planning efficiency;

[0085] (5) The present invention optimizes the primary and secondary road networks through an adaptive smoothing algorithm, significantly improving the practicality and safety of the road network, so that the generated road network can naturally fit the terrain changes in complex terrain. Through the optimization of intersections based on direction vectors, unreasonable designs such as sharp turns are reduced, the smoothness of the connection between secondary roads and primary roads is improved, and the rationality and smoothness of the generated road network are greatly improved;

[0086] (6) The present invention generates a visual three-dimensional model by inputting the generated road network into the smart city design platform, and uses VR\AR technology to assist in dynamic interactive adjustment and improvement of the generated plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0087] Figure 1 is a flow chart of the present invention;

[0088] Figure 2 This is a schematic diagram of the principle of optimizing the road angle in the present invention. DETAILED DESCRIPTION

[0089] The technical solution of the present invention is further described below in conjunction with the accompanying drawings.

[0090] The technical solution of the present invention will be described in detail below by taking the urban design of a certain area in Chongqing as an example.

[0091] like Figure 1 As shown, a method for intelligently generating a three-dimensional urban road network for complex mountainous terrain includes the following steps:

[0092] S1. Construction of design site database: Collect the current terrain data and environmental data of the design site, collect the relevant upper-level planning data of the design site, import them into the geographic information platform, process them into raster data or vector data in a unified format, and generate the design site database;

[0093] S1 specifically includes the following steps:

[0094] S1.1 Collect the current terrain data of the design site and obtain DEM (Digital Elevation Model) raster data through the geospatial data cloud or the geological survey bureau of the city;

[0095] S1.2 Collect the current environmental data of the design site, including vector environmental data of rivers and current buildings, atmospheric environmental data and raster data of land cover types;

[0096] Obtain vector environmental data of rivers and existing buildings through open source data platforms; obtain atmospheric environmental data through the China Meteorological Data Network; obtain raster data of land cover types through the National Geographic Information Resource Directory Service System;

[0097] S1.3 Use drones equipped with high-resolution cameras and LiDAR to take aerial images of the design site, and identify the current terrain data of the design site and the current environmental data of the design site (except for the atmospheric environment data) from the aerial images to assist in improving the current terrain data and environmental data of the design site;

[0098] S1.4 Obtain the superior planning data of the design site from the local urban planning department. The superior planning data includes the city master plan, regional development plan, transportation plan and superior planning cultivated land indicators;

[0099] S1.5 Import the collected terrain data, environmental data and superordinate planning data into a high-performance workstation equipped with a multi-core CPU and large-capacity memory, with at least 64GB RAM. Use the geographic information platform to select a coordinate system based on the site coordinates, add geographic registration control points, and spatially locate the superordinate planning data. In the geographic information platform, convert the planning map of the superordinate planning data from raster data to vector elements, correct errors in the conversion process, identify the scope of various functional areas in the planning map, assign values ​​to different functional areas, and store them as raster data.

[0100] S1.6 Integrate the data to generate the design site database, store it in the RAID 10 storage array, and use the database manager to manage the data;

[0101] S2. Identification of trunk road nodes based on area development suitability: Calculate the development suitability of each area in the design site based on the design site database, and identify the center point of each area as the trunk road node in the high suitability area after threshold screening;

[0102] S2 specifically includes the following steps:

[0103] S2.1 Based on the design site database, the design site is analyzed using the geographic information platform to obtain the evaluation results of the normalized slope index, regional climate comfort index, land transformation index, and functional area transformation time series index;

[0104] The slope was calculated using DEM raster data, and the slope was evaluated. It was divided into five levels: high, relatively high, medium, relatively low, and low, with scores of 1, 2, 3, 4, and 5, according to ≤2°, 2-6°, 6-15°, 15-25°, and >25°. The high level corresponds to flat land, the relatively high level corresponds to flat slope land, the medium level corresponds to gentle slope land, the relatively low level corresponds to gentle steep slope land, and the relatively low level corresponds to steep slope land. A slope classification map was generated.

[0105] The atmospheric environment data is used to calculate the average monthly temperature and the average monthly relative humidity, calculate the temperature and humidity index, and evaluate the regional climate comfort level. The regional climate comfort level is divided into 7 levels and scored according to the level. The temperature and humidity index of 60-65 corresponds to the comfort level of 7, with 7 points; the temperature and humidity index of 56-60 or 65-70 corresponds to the comfort level of 6, with 6 points; the temperature and humidity index of 50-56 or 70-75 corresponds to the comfort level of 5, with 5 points; the temperature and humidity index of 45-50 or 75-80 corresponds to the comfort level of 4, with 4 points; the temperature and humidity index of 40-45 or 80-85 corresponds to the comfort level of 3, with 3 points; the temperature and humidity index of 32-40 or 85-90 corresponds to the comfort level of 2, with 2 points; the temperature and humidity index <32 or >90 corresponds to the comfort level of 1, with 1 point; the higher the comfort level, the higher the comfort;

[0106] Using land cover type data and based on the upper-level planned cultivated land indicators, the degree of land transformation is evaluated, and each type of land is divided into three levels: transformable, moderately transformable, and non-transformable, and each is assigned a score of 5, 3, and 1.

[0107] Using the city master plan, each functional area is divided into three levels: pending development, near-term development and key development, and is scored 1, 2 and 3 points;

[0108] The four evaluation scores were normalized to obtain the slope index, regional climate comfort index, land transformation index, and functional area transformation time series index;

[0109] S2.2 According to the analytic hierarchy process (AHP), a judgment matrix is ​​constructed, and the importance of the matrix is ​​ranked as follows: slope index > land transformability index > functional area transformation time series index > regional climate comfort index; calculate the sum of each column of the judgment matrix; divide each element by the sum of its column to obtain a normalized judgment matrix; then calculate the average value of each row to obtain the weight of each factor;

[0110] The slope index, regional climate comfort index, land transformability index, and functional area transformation time series index are superimposed and calculated according to the weights calculated by the judgment matrix to obtain the basic map of development suitability of the design site and store it in raster data format;

[0111] The calculation formula for the superposition of development suitability is: k1×w1+k2×w2+k3×w3+k4×w4, where k1 is the slope index, w1 is the weight of the slope index, k2 is the regional climate comfort index, w2 is the weight of the regional climate comfort index; k3 is the land transformation index, w3 is the weight of the land transformation index; k4 is the functional area transformation time series index, w4 is the weight of the functional area transformation time series index;

[0112] S2.3 Establish a 1000m grid and use the Raster to Point tool to convert the development suitability base map into point data; use the Spatial Join tool to extract the point data values ​​to the corresponding grid, and use the natural segmentation method to classify the extracted data into 5 levels, from 1 to 5 corresponding to low to high development suitability, and store the classification results in a new field of "Development Suitability"; use the Select by Attribute tool in the attribute table to filter out cells with a value of 5 in the Development Suitability field, i.e. cells with high development suitability, and take the geometric center of each high suitability area as the main road node;

[0113] S2.4 Check whether the location of each node has steep slopes or terrain that is not suitable for development. If so, move the node to the nearest flat area;

[0114] S3. Generation of trunk road network based on improved A* algorithm: Generate a three-dimensional grid of the design site, use the improved A* algorithm to search for paths, connect the trunk road nodes and extend them to the edge of the design site to form several trunk roads, delete unreasonable paths, calculate the distance between trunk roads, and make supplementary adjustments to the road network;

[0115] S3 specifically includes the following steps:

[0116] S3.1 identifies the center point of each grid of the design site DEM grid data containing height coordinates, connects all grid center points to generate a three-dimensional grid, writes the three-dimensional grid using a high-speed storage device NVMe SSD, and uses it to generate road network storage;

[0117] S3.2 uses the improved A* algorithm to search for paths, defines the main road nodes as the starting point and end point of the path search, and the formula of the function f(n) of the improved A* algorithm is:

[0118] f(n)=g(n)+w·h(n)+α·S(n)+β·C(n)

[0119] g(n) is the actual cost from the starting point to the current node n, that is, the path length;

[0120] h(n) is the Euclidean distance heuristic function, which estimates the cost from node n to the target node;

[0121] w is a dynamic weight, which is used to adjust the influence of h(n) and changes dynamically according to the progress of the search;

[0122] S(n) is the terrain cost function, which reflects the slope or terrain undulation of the area where node n is located;

[0123] α is the weight of the terrain cost, which determines the impact of terrain on path selection;

[0124] C(n) is the cost function of the road design rule, which takes into account the factors of road spacing and road curvature;

[0125] β is the weight of the road design rule, which determines the impact of the road design rule on path selection;

[0126] in,

[0127] g(n)=g(p)+d(p,n)

[0128] g(p) is the cost of the previous node p;

[0129] d(p,n) is the path length from node p to node n;

[0130]

[0131] Δh(n) is the height difference from the previous node p to the current node n, i.e. the slope;

[0132] d(p,n) is the horizontal distance from the previous node p to the current node n;

[0133]

[0134] (x n ,y n ,z n ) is the three-dimensional coordinate of the current node n;

[0135] (x goal ,y goal ,z goal) is the three-dimensional coordinate of the target node;

[0136]

[0137] d road (n) is the actual distance between the road at the current node n and the nearest adjacent road;

[0138] d target The ideal road spacing is predefined, which is 800-1000 meters according to urban design specifications;

[0139] θ(n) is the turning angle of the current path at node n, expressed in degrees;

[0140] θmax is the maximum allowed road turning angle, which is determined to be 60 degrees according to the design standard;

[0141] S3.3 Use the improved A* algorithm in step S3.2 to perform path search, define the trunk road node as the starting point of the path search, design the site three-dimensional grid boundary point as the end point of the path search, generate a path with the minimum f(n) connected to the grid boundary from each trunk road node n, and preliminarily form a trunk road network;

[0142] S3.4 checks the node degree of the main road nodes after the road network is generated, identifies and prioritizes all paths connected to nodes whose node degrees exceed the expected value of 4, with the paths connecting the grid boundaries having the highest priority, and the remaining paths being ranked based on the cost function, which comprehensively considers the slope and the distance, and deletes the path with the lowest priority until the node degree is less than or equal to 4;

[0143] The cost function formula is:

[0144]

[0145] w1 and w2 are weight coefficients, which control the influence of slope cost and distance cost respectively;

[0146] Δh r is the height difference between the starting point and the end point of path r;

[0147] d r is the length of path r;

[0148] S3.5 calculates the spacing between the generated trunk road networks, and marks all areas with spacing greater than 1200 meters according to the trunk road spacing standard; in each marked area, calculates its geometric center point as a newly added node, and uses the improved A* algorithm in step S3.2 to search for a path from the node to the existing trunk road node with the longest distance among all trunk roads at the edge of the area to obtain a connecting path; extends the trunk roads from both ends of the connecting path to the three-dimensional grid boundary of the design site to form several trunk roads;

[0149] S4. Generation of secondary trunk road network based on improved A* algorithm: Based on the generated trunk road network, according to the road spacing specification, secondary trunk road nodes are selected on the trunk road, and the improved A* algorithm is used to perform path search to generate the secondary trunk road network;

[0150] S4 specifically includes the following steps:

[0151] S4.1 Select secondary trunk road nodes on the generated trunk road network according to the spacing rule mark, the spacing rule is greater than 350 meters and less than 500 meters, and the closer to the three-dimensional grid boundary, the larger the spacing;

[0152] S4.2 For each area divided by the trunk road network, use the improved A* algorithm in step S3.2 to search for a path, adjust the function formula of the improved A* algorithm, increase the weight of the road design rule cost, that is, increase the β value in the formula, define the secondary trunk road node on the boundary of the area as the starting point and the end point, and the nodes on the same trunk road path cannot be used as a pair of starting points and end points;

[0153] S4.3 Check the node degree of the secondary trunk road nodes after the path is generated, identify the nodes whose degree exceeds the expected value of 1, do not calculate the main road network paths and paths outside the local area, delete the redundant paths so that the degree of all nodes is less than or equal to 1, and obtain the secondary trunk road network;

[0154] S5. Smoothing of primary and secondary road networks and optimization of road angles: Use an adaptive smoothing algorithm to process the generated primary and secondary road networks and optimize road angles;

[0155] S5 specifically includes the following steps:

[0156] S5.1 Divide the road network into different areas. According to the terrain characteristics, the road network is divided into flat areas with a slope of 0°-5°, hilly areas with a slope of 5°-15°, and mountainous areas with a slope of more than 15°;

[0157] S5.2 dynamically selects and adjusts the smoothing algorithm and window according to different regions. The window represents the number of adjacent data points selected within a given range. For flat areas, the linear interpolation method is selected, and the parameters use a larger smoothing window; for hilly areas, the Bezier curve is selected, and the parameters use a medium-sized smoothing window; for mountainous areas, the Gaussian filter is used, and the parameters use a smaller smoothing window;

[0158] S5.3 converts the smoothed road network into a multidimensional vector data set, calculates the direction vector using Python, and calculates the angle based on the direction vector;

[0159] S5.4 Identify the intersection node A with an angle less than 70 degrees, and divide the two direction vectors involved in the angle into the main section and the secondary section; take the node as the starting point, and generate an alternative node B on the main section with a step length of cot(θ)*10m; generate a new section perpendicular to the main section from the alternative node B, and calculate the intersection C of the new section and the secondary section; recalculate the local direction vector and calculate the angle, and repeat the calculation until the road angle is greater than 70 degrees, such as Figure 2 As shown;

[0160] S5.5 writes the generated road network into a high-speed storage device;

[0161] S6. Visualization output and interactive optimization: Output the generated road network and perform 3D visualization interactive optimization;

[0162] S6 specifically includes the following steps:

[0163] S6.1 inputs the generated road network data into the virtual reality application software on the development workstation equipped with high-performance CPU and GPU for processing and rendering;

[0164] S6.2 connects to VR headsets and supporting controllers, integrates and tests different VR devices, and creates VR scenes; connects to AR devices, creates AR scenes, and places and tracks virtual objects through mobile devices or AR glasses; performs 3D visualization and interactive optimization of the generated road network on 4K and above high-resolution displays, and uses VR and AR tracking devices to accurately adjust and optimize the scene.

Claims

1. A method for intelligently generating a three-dimensional urban road network for complex mountainous terrain, characterized in that: The steps include: (1) Collect the current terrain data and environmental data of the design site, collect the upper-level planning data of the design site, import them into the geographic information platform, process them into raster data or vector data in a unified format, and generate the design site database; (2) Based on the design site database, the development suitability of each area of ​​the design site is calculated, and after threshold screening, the center point of each area is identified as the main road node in the high suitability area; (3) Generate a three-dimensional grid of the design site, use the improved A* algorithm to search for paths, connect the main road nodes and extend them to the edge of the design site to form several main roads, delete unreasonable paths, calculate the distance between the main roads, make supplementary adjustments to the road network, and generate the main road network; (4) Based on the generated trunk road network, secondary trunk road nodes are selected on the trunk road according to the road spacing specification, and the improved A* algorithm is used to perform path search to generate the secondary trunk road network; (5) Use an adaptive smoothing algorithm to process the generated primary and secondary road networks and optimize the road angles; (6) Output the generated road network and perform three-dimensional visualization interactive optimization.

2. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 1, characterized in that: The step (1) specifically comprises the following steps: (1.1) Collect the current terrain data of the design site and obtain DEM raster data; (1.2) Collect the current environmental data of the design site, including vector environmental data of rivers and current buildings, atmospheric environmental data and raster data of land cover types; (1.3) Take aerial images of the design site, identify the current terrain data of the design site and the current environmental data of the design site (excluding atmospheric environmental data) from the aerial images, and use them to assist in improving the current terrain data and environmental data of the design site; (1.4) Obtain the superior planning data of the design site, which includes the city master plan, regional development plan, transportation plan and superior planning cultivated land indicators; (1.5) Import the collected terrain data, environmental data and superordinate planning data into a high-performance workstation, use the geographic information platform, select the coordinate system based on the site coordinates, add georeferencing control points, and spatially locate the superordinate planning data; In the geographic information platform, the planning map in the upper planning data is converted from raster data to vector elements, errors in the conversion process are corrected, the scope of various functional areas in the planning map is identified, different functional areas are assigned values, and stored as raster data; (1.6) Integrate data to generate a design site database, store and manage data using a database manager.

3. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 2, characterized in that: In the step (1.2), vector environmental data of rivers and existing buildings are obtained through an open source data platform, atmospheric environmental data are obtained through the China Meteorological Data Network, and raster data of surface coverage types are obtained through the National Geographic Information Resource Directory Service System.

4. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 2, characterized in that: The step (2) specifically comprises the following steps: (2.1) Based on the design site database, the design site is analyzed using the geographic information platform to obtain the evaluation results of the normalized slope index, regional climate comfort index, land transformation index, and functional area transformation time series index; (2.2) According to the analytic hierarchy process (AHP), a judgment matrix is ​​constructed to obtain the weight of each factor; The slope index, regional climate comfort index, land transformability index, and functional area transformation time series index are superimposed and calculated according to the weights calculated by the judgment matrix to obtain the basic map of development suitability of the design site and store it in raster data format; (2.3) Establish a grid, convert the development suitability base map into point data, use the spatial connection tool to extract the value of the point data into the corresponding grid, and use the natural segmentation method to classify the extracted data into 5 levels, from 1 to 5 corresponding to low to high development suitability, and store the classification results in a new field "Development Suitability"; use the attribute selection tool in the attribute table to filter out the cells with a value of 5 in the development suitability field, that is, cells with high development suitability, and take the geometric center of each high suitability area as the main road node; (2.4) Check whether the location of each node has steep slopes or terrain that is not suitable for development. If so, move the node to the nearest flat area.

5. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 4, characterized in that: The step (2.1) specifically comprises the following steps: The slope was calculated using DEM raster data, and the slope was evaluated. It was divided into five levels: high, relatively high, medium, relatively low, and low, with scores of 1, 2, 3, 4, and 5, according to ≤2°, 2-6°, 6-15°, 15-25°, and >25°. The high level corresponds to flat land, the relatively high level corresponds to flat slope land, the medium level corresponds to gentle slope land, the relatively low level corresponds to gentle steep slope land, and the relatively low level corresponds to steep slope land. A slope classification map was generated. The atmospheric environment data is used to calculate the average monthly temperature and the average monthly relative humidity, and the temperature and humidity index is calculated to evaluate the regional climate comfort level. The regional climate comfort level is divided into 7 levels and scored according to the level. The temperature and humidity index of 60-65 corresponds to the comfort level of 7, with 7 points; the temperature and humidity index of 56-60 or 65-70 corresponds to the comfort level of 6, with 6 points; the temperature and humidity index of 50-56 or 70-75 corresponds to the comfort level of 5, with 5 points; the temperature and humidity index of 45-50 or 75-80 corresponds to the comfort level of 4, with 4 points; the temperature and humidity index of 40-45 or 80-85 corresponds to the comfort level of 3, with 3 points; the temperature and humidity index of 32-40 or 85-90 corresponds to the comfort level of 2, with 2 points; the temperature and humidity index <32 or >90 corresponds to the comfort level of 1, with 1 point; the higher the comfort level, the higher the comfort; Using land cover type data and based on the upper-level planned cultivated land indicators, the degree of land transformation is evaluated, and each type of land is divided into three levels: transformable, moderately transformable, and non-transformable, and each is assigned a score of 5, 3, and 1. Using the city master plan, each functional area is divided into three levels: pending development, near-term development and key development, and is scored 1, 2 and 3 points; The four evaluation scores were normalized to obtain the slope index, regional climate comfort index, land transformability index, and functional area transformation time series index.

6. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 4, characterized in that: The order of importance of the judgment matrix in step (2.2) is: slope index > land transformability index > functional area transformation time series index > regional climate comfort index; Calculate the sum of each column of the judgment matrix; divide each element by its column sum to obtain a normalized judgment matrix; then calculate the average value of each row to obtain the weight of each factor; The superposition calculation formula for development suitability is: k1×w1+k2×w2+k3×w3+k4×w4, where k1 is the slope index, w1 is the weight of the slope index, k2 is the regional climate comfort index, w2 is the weight of the regional climate comfort index; k3 is the land transformability index, w3 is the weight of the land transformability index; k4 is the functional area transformation timing index, and w4 is the weight of the functional area transformation timing index.

7. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 4, characterized in that: The step (3) specifically comprises the following steps: (3.1) Identify the center point of each grid of the design site DEM grid data containing height coordinates, connect all grid center points to generate a three-dimensional grid, write the three-dimensional grid using a high-speed storage device, and use it to generate a road network storage; (3.2) Use the improved A* algorithm to search for paths, define the main road nodes as the starting point and end point of the path search, and the formula of the function f(n) of the improved A* algorithm is: f(n)=g(n)+w·h(n)+α·S(n)+β·C(n) g(n) is the actual cost from the starting point to the current node n, that is, the path length; h(n) is the Euclidean distance heuristic function, which estimates the cost from node n to the target node; w is a dynamic weight, which is used to adjust the influence of h(n) and changes dynamically according to the progress of the search; S(n) is the terrain cost function, which reflects the slope or terrain undulation of the area where node n is located; α is the weight of the terrain cost, which determines the impact of terrain on path selection; C(n) is the cost function of the road design rule, which takes into account the factors of road spacing and road curvature; β is the weight of the road design rule, which determines the impact of the road design rule on path selection; in, g(n)=g(p)+d(p,n) g(p) is the cost of the previous node p; d(p,n) is the path length from node p to node n; Δh(n) is the height difference from the previous node p to the current node n, i.e. the slope; d(p,n) is the horizontal distance from the previous node p to the current node n; (x n ,y n ,z n ) is the three-dimensional coordinate of the current node n; (x goal ,y goal ,z goal ) is the three-dimensional coordinate of the target node; d road (n) is the actual distance between the road at the current node n and the nearest adjacent road; d target is a predefined ideal road spacing; θ(n) is the turning angle of the current path at node n, expressed in degrees; θ max is the maximum allowed road turning angle; (3.3) Use the improved A* algorithm in step (3.2) to perform path search, define the trunk road node as the starting point of the path search, design the site 3D grid boundary point as the end point of the path search, generate a path with the minimum f(n) from each trunk road node n to the grid boundary, and preliminarily form a trunk road network; (3.4) Check the node degree of the main road nodes after the road network is generated, identify all the paths connected to the nodes whose degree exceeds the expected value of 4, and prioritize them. The path connecting the grid boundary has the highest priority, and the remaining paths are ranked based on the cost function. The cost function comprehensively considers the slope and distance, and deletes the path with the lowest priority until the node degree is less than or equal to 4; The cost function formula is: w1 and w2 are weight coefficients, which control the influence of slope cost and distance cost respectively; Δh r is the height difference between the starting point and the end point of path r; d r is the length of path r; (3.5) Calculate the spacing between the generated trunk road networks, and mark all areas with a spacing greater than 1200 meters according to the trunk road spacing standard; in each marked area, calculate its geometric center point as a new node, and use the improved A* algorithm in step (3.2) to search for a path from this node to the existing trunk road node with the longest distance among all trunk roads at the edge of the area to obtain a connecting path; extend the trunk roads from both ends of the connecting path to the three-dimensional grid boundary of the design site to form several trunk roads.

8. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 7, characterized in that: The step (4) specifically comprises the following steps: (4.1) Select secondary trunk road nodes on the generated trunk road network according to the spacing rule. The spacing rule is greater than 350 meters and less than 500 meters, and the closer to the three-dimensional grid boundary, the larger the spacing; (4.2) For each area divided by the trunk road network, use the improved A* algorithm in step (3.2) to search for paths, adjust the function formula of the improved A* algorithm, increase the weight of the road design rule cost, that is, increase the β value in the formula, define the secondary trunk road nodes on the boundary of the area as the starting point and end point, and the nodes on the same trunk road path cannot be used as a pair of starting points and end points; (4.3) After the path is generated, the node degree of the secondary trunk road nodes is checked, and the nodes with node degrees exceeding the expected value of 1 are identified. The main road network paths and paths outside the area are not calculated, and the redundant paths are deleted so that the node degrees of all nodes are less than or equal to 1, and the secondary trunk road network is obtained.

9. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 8, characterized in that: The step (5) specifically comprises the following steps: (5.1) Divide the road network into different areas. According to the terrain characteristics, the road network is divided into flat areas with a slope of 0°-5°, hilly areas with a slope of 5°-15°, and mountainous areas with a slope of more than 15°; (5.2) Dynamically select and adjust the smoothing algorithm and window according to different regions. The window represents the number of adjacent data points selected within a given range. For flat areas, the linear interpolation method is selected, and the parameters use a larger smoothing window; for hilly areas, the Bezier curve is selected, and the parameters use a medium-sized smoothing window; for mountainous areas, the Gaussian filter is used, and the parameters use a smaller smoothing window; (5.3) Convert the smoothed road network into a multidimensional vector data set, use Python to calculate the direction vector, and calculate the angle based on the direction vector; (5.4) Identify the intersection node A where the angle is less than 70 degrees, and divide the two direction vectors involved in the angle into the main section and the secondary section; Taking this node as the starting point, generate an alternative node B on the main section with a step length of cot(θ)*10m; generate a new section perpendicular to the main section from the alternative node B, and calculate the intersection C of the new section and the secondary section; Recalculate the local direction vector and the angle, and repeat the calculation until the road angle is greater than 70 degrees; (5.5) Write the generated road network into a high-speed storage device.

10. The method for intelligently generating a three-dimensional urban road network for complex mountainous terrain according to claim 9, characterized in that: The step (6) specifically comprises the following steps: (6.1) Input the generated road network data into the virtual reality application software on the development workstation for processing and rendering; (6.2) Connect VR headsets and supporting controllers, integrate and test different VR devices, and create VR scenes; Connect AR devices and place and track virtual objects through mobile devices or AR glasses; interactively optimize the generated 3D visualization of the road network on the display, and use VR and AR tracking devices to adjust and optimize the scene.

Images