Simulation road network generation method for realizing fine optimization of ramp

Abstract

This invention discloses a simulation road network generation method for achieving fine-grained optimization of ramps, belonging to the field of traffic simulation. It aims to address industry pain points such as low ramp fidelity, distorted lane count statistics for ramps with high curvature, and low efficiency of manual correction for large-scale road networks in existing OSM-based simulation road networks. This method first acquires the regional OSM road network and performs format conversion to obtain the initial road network file for the target region. Then, it uses high-precision SHP files to filter ramp candidate data and construct a dedicated layer, accurately generating ramp centerlines and actual boundary lines to lock the ramp area. Subsequently, it performs grid-based cropping of the lane line images in the ramp area, and uses bidirectional auxiliary lines to cut and count the number of intersections with lane lines, accurately calculating the number of lanes and lane widths in each segment. Finally, it uses the aforementioned precise parameters to correct the initial road network. This invention can avoid lane count statistics errors for ramps with high curvature, achieve automated and refined modeling of all ramp elements, significantly improve the fidelity and construction efficiency of traffic simulation road networks, and adapt to the needs of batch processing of large-scale road networks.

Description

Technical Field

[0001] This invention relates to the field of traffic simulation, and in particular to a method for generating a simulated road network that achieves fine optimization of ramps. Background Technology

[0002] The purpose of traffic simulation is to reproduce or predict the traffic operation of existing or future systems, thereby explaining and analyzing complex traffic phenomena, identifying the root causes of problems, and ultimately optimizing the traffic system under study. Microscopic simulation models use the most basic elements of the traffic system, such as individual vehicles, lanes, and traffic lights, as modeling units, thus accurately and flexibly reflecting the impact of various road and traffic conditions. However, the realism and accuracy of the simulation largely depend on the accuracy of the underlying road network; the difference between the simulated traffic network and the real road network significantly affects the simulation's effectiveness. While existing traffic simulations can construct large-scale road networks, their highway ramp characteristics and lane numbers are generated automatically by their underlying logic, resulting in inconsistencies between ramp features and lanes and real-world conditions.

[0003] In engineering applications, shapefiles are widely used to provide road network data. This type of GIS data supports various geometric types, such as points, lines, and polygons, and can store a large amount of spatial information, including geometric coordinates and topological relationships, supporting high-precision GIS analysis and simulation. Existing microscopic traffic simulations mostly rely on crawling OSM data to obtain net.xml files. While this method can construct large-scale road networks, it lacks refinement and timeliness. Real highway ramp characteristics or road quantities are not updated in the OSM in a timely manner, and the crawled road network structure data has a delay.

[0004] In terms of road network construction, traffic elements such as highway ramp features and lanes are simply built using the underlying logic of simulation, which cannot realistically reproduce the characteristics of real-world ramps. Although simulation can support manual modification of ramp features, manual modification is extremely inefficient for large-scale road networks. The starting point of this patent is to explore how to automatically correct existing basic road networks using existing high-precision SHP data. This invention aims to develop a technology that can integrate high-precision SHP data with basic road network information to generate a refined simulated road network.

[0005] Some existing technologies address the problems of current traffic simulation relying on OSM data to construct road networks, insufficient refinement, and lagging updates. They propose methods based on high-precision SHP data, through a series of data processing and optimization, to achieve automated construction of two-dimensional and three-dimensional simulated road networks. However, these methods have shortcomings such as the inability to accurately reproduce ramp morphology and lane parameters, and low efficiency and large errors in manual correction of large-scale road networks. Summary of the Invention

[0006] The inventors discovered that in the prior art, the number of lanes is generally determined by finding the intersection points of lane lines with equidistant linear arrays parallel to the x-axis or y-axis. However, when this method is applied to ramps, the large curvature of the ramps leads to a larger number of intersection points than the actual number of lane lines. Therefore, the purpose of this invention is to provide a simulation road network generation method for achieving fine optimization of ramps, thereby overcoming the shortcomings of the prior art.

[0007] The objective of this invention can be achieved through the following technical solutions: A method for generating a simulated road network to achieve fine-grained optimization of ramps, comprising: Step S1: Obtain the regional OSM road network and perform format conversion to obtain the initial road network file containing the target region; Step S2: Obtain the basic geographic information SHP file of the target area, read the full road network layer information and filter out all ramp candidate data to construct a ramp layer, generate the center line and actual boundary line of the ramp based on the constructed ramp layer, and further obtain the ramp area; Step S3: Based on the basic geographic information SHP file, read the lane line layer, generate the first image corresponding to the ramp area, and crop the first image to obtain multiple first sub-images; Step S4: Cut each first sub-image using a first auxiliary line parallel to the x-axis and / or a second auxiliary line parallel to the y-axis to obtain the number of intersections between each first auxiliary line and / or second auxiliary line and the lane line; Step S5: Obtain the number of lane lines corresponding to the corresponding first sub-image based on the number of intersections between each first auxiliary line and / or second auxiliary line and the lane line; Step S6: Based on the number of lane lines in each first sub-image, obtain the number of lanes and lane width within the ramp; Step S7: Based on the obtained ramp area and the number and width of lanes within the ramp, modify the initial road network file.

[0008] Step S2 includes: Step S2-1: Obtain the basic geographic information SHP file of the target area, read the full road network layer information and filter out all ramp candidate data to construct the ramp layer; Step S2-2: Obtain the latitude and longitude coordinate sequence of the road centerline within the ramp layer, and fit it to obtain the ramp centerline; Step S2-3: Based on the preset ramp width parameters, extend symmetrically to both sides with the center line as the reference to generate the original boundary line of the ramp; Step S2-4: Intersect the original boundary line of the ramp with the boundary lines of other roads to determine the actual boundary line of the ramp, and further obtain the ramp area.

[0009] The filtering rules for obtaining candidate ramp data in step S2-1 include: Rule 1: The topological relationship only connects the main line of the highway and the interchange nodes, with no other non-interchange type connected objects; Rule 2: The geometric shape consists of short linear segments with a clear curve or broken line orientation; Rule 3: The road network unit to which the road segment belongs has a direct coding association with the interchange and the main line of the expressway.

[0010] The specific examples of obvious curves or broken lines in Rule 2 are as follows: There exists a road whose center point has a tangent angle that is less than the first threshold angle, and whose vector distance is less than the first threshold angle.

[0011] Step S3 includes: Step S3-1: Read the lane line layer based on the basic geographic information SHP file and generate the first image corresponding to the ramp area; Step S3-2: Determine the grid size based on the preset ramp width parameters; Step S3-3: Perform atmospheric treatment on the first image based on the grid size to obtain multiple candidate sub-images; Step S3-4: Based on the area ratio of the ramp region in all candidate sub-images, select a portion of candidate sub-images as the first sub-image.

[0012] Step S3-3 includes: Step S3-3-1: Sample multiple first sampling points are obtained by sampling along the centerline of the ramp; Step S3-3-2: Using the first sampling point as the center and the grid size as the side length, crop the first image to obtain multiple candidate sub-images.

[0013] In step S4, multiple first auxiliary lines and multiple second auxiliary lines are provided.

[0014] When using the first auxiliary line and the second auxiliary line for cutting, in step S5, the maximum value of the number of intersections of all auxiliary lines is taken as the number of lane lines.

[0015] A simulated road network generation device for achieving fine-grained optimization of ramps includes a memory, a processor, and a program stored in the memory, wherein the processor executes the program to implement the method described above.

[0016] A storage medium having a program stored thereon, which, when executed, implements the method described above.

[0017] Compared with the prior art, the present invention has the following beneficial effects: 1. By first determining the ramp area, then constructing the first image of the ramp area based on the basic geographic information SHP file, and then cropping the first image to obtain multiple first sub-images, and then calculating the intersection points of the lanes in each first sub-image, the problem of the number of intersection points being greater than the actual number of lane lines caused by the same auxiliary line intersecting with different sections of the ramp is avoided.

[0018] 2. A closed-loop processing logic for the entire ramp area, from data screening to boundary locking, was constructed. The generation of the ramp area was broken down into four progressive steps: layer construction, centerline fitting, original boundary generation, and actual boundary determination. This achieved accurate restoration of the ramp spatial range and solved the core problems of ambiguous ramp boundaries and disconnection from the surrounding road network topology in existing technologies. First, a dedicated ramp layer was used to decouple the ramp from the ordinary road network, reducing redundant calculations in the overall road network processing and defining a dedicated range for subsequent lane parameter extraction, significantly reducing interference from the surrounding road network on the refined processing of the ramp. Boundaries were generated based on the centerline, and then corrected by intersections with other road boundaries to obtain the actual boundaries. This avoided the problem of boundary conflicts between the preset ramp width and the real road network, ensuring a high degree of matching between the ramp geometry and the real road network, and improving the accuracy of the simulated road network.

[0019] 3. A three-dimensional ramp screening system integrating topological relationships, geometric morphology, and coding association was established. These three rules form a complementary verification mechanism, locking in ramp data from three dimensions: functional positioning, morphological characteristics, and attribute association. This significantly improves the accuracy of ramp identification and prevents the mixing of non-ramp data such as highway mainlines and ordinary roads from the source, ensuring the purity of the ramp layer. The standardized screening rules can be fully embedded into automated processing workflows without manual intervention, solving the pain points of low efficiency and high misjudgment rate of manual ramp screening in large-scale road networks. It perfectly adapts to the batch processing needs of large-scale road networks. The core topological features of ramps are strictly limited (only connecting highway mainlines and interchange nodes), accurately matching the traffic function attributes of ramps and avoiding misidentification of other connecting lines and auxiliary road data, laying a solid data foundation for subsequent refined optimization of ramps.

[0020] 4. The qualitative description of obvious curves or broken lines is transformed into a quantifiable and calculable tangent angle determination index, solving the subjectivity problem of manual identification and achieving automated and standardized determination of ramp morphology characteristics. This ensures consistency in judgment across different regions and batches of road networks. Based on the determination method using the tangent angle at the road center point, it exhibits extremely high sensitivity in identifying ramps with small radius curves and continuous broken line segments. It avoids missing irregularly shaped ramps and misjudging smooth straight road sections, further narrowing the boundaries of ramp screening and improving identification accuracy. By adjusting distance and angle thresholds, it can flexibly adapt to road network scenarios with different design standards, such as urban expressway ramps, mountain highway ramps, and hub interchange ramps, significantly improving the method's compatibility and applicability.

[0021] 5. A gridded cropping and effective filtering mechanism for ramp lane line images was established. This mechanism decomposes the complete high-curvature ramp area into multiple standardized sub-image units, fundamentally solving the core problem of inaccurate lane count statistics caused by repeated intersections of a single auxiliary line with multiple lane lines when processing long curved ramps as a whole. The grid size is matched based on the ramp width parameter, balancing computational accuracy and processing efficiency. This avoids both excessively large grids leading to high ramp curvature and statistical inaccuracies within sub-images, and excessively small grids causing a surge in computational load. Effective sub-images are filtered by ramp area proportion, eliminating redundant background areas and retaining only core effective road segments for subsequent processing. This significantly reduces invalid computation and avoids the impact of background interference on lane line recognition, further improving the accuracy of lane count statistics.

[0022] 6. Cropping is performed centered on the sampling point of the ramp centerline, ensuring that the ramp lane lines in each sub-image are always within the core area of ​​the image. This avoids recognition and statistical errors caused by lane lines being cut off or deviating from the image, significantly improving the reliability of subsequent lane line intersection counts. The cropping method, which continuously samples along the ramp centerline, achieves seamless and equidistant coverage of the entire ramp segment. Whether it's a straight or curved ramp, complete segment processing is achieved without any omissions, ensuring the integrity of lane parameter extraction for the entire ramp. The center-cropping mode makes the ramp shape more regular in each sub-image, significantly reducing the ramp curvature within the sub-image. This fundamentally alleviates the distortion problem in auxiliary line intersection counts caused by ramps with high curvature, providing key support for accurate ramp lane number counting.

[0023] 7. Multiple parallel auxiliary lines are used for cross-cutting, and redundancy verification is achieved through multi-line statistics. This avoids statistical errors caused by missing lane lines, occlusion, or noise interference from a single auxiliary line, significantly improving the robustness of lane line count. Multiple auxiliary lines can completely cover the entire lane line area within the sub-image, preventing statistical omissions caused by a single auxiliary line falling precisely on lane line gaps or avoiding bends. This design can adapt to ramp lane lines at different angles and directions within the sub-image, capturing more comprehensive lane line information. The multi-auxiliary-line design provides a sufficient data foundation for subsequent statistical result screening and calibration, a prerequisite for accurate lane count determination.

[0024] 8. The rule of using the maximum value of superimposed auxiliary lines in both directions perfectly adapts to any orientation of ramp lane lines: when lane lines are nearly horizontal, the y-axis parallel auxiliary line statistics are more accurate; when lane lines are nearly vertical, the x-axis parallel auxiliary line statistics are more accurate. Taking the maximum value ensures that the number of lane lines most closely matches reality, solving the problem of statistical distortion for inclined / bent lane lines caused by single-direction auxiliary lines. The maximum value statistical rule effectively avoids the problem of insufficient intersection counts caused by local breaks, contamination, or blurring of lane lines, maximizing the restoration of the true total number of lane lines and avoiding the error of underestimating lane counts caused by statistical methods such as mean and mode. This rule has simple logic and extremely high computational efficiency, requiring no complex algorithm model for implementation. While ensuring statistical accuracy, it also meets the computational power requirements of large-scale road network batch processing, perfectly adapting to engineering application scenarios of automated simulated road network generation. Attached Figure Description

[0025] Figure 1 This is a schematic diagram of the main steps of the method of the present invention; Figure 2 A diagram illustrating an example of extracting the number of lanes; Figure 3 A schematic diagram illustrating an example of extracting road width; Figure 4 This is a schematic diagram of the modified road network. Detailed Implementation

[0026] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0027] A method for generating a simulated road network that achieves fine-grained optimization of ramps, such as Figure 1 As shown, it includes: Step S1: Obtain the regional OSM road network and perform format conversion to obtain the initial road network file containing the target region; In this embodiment, OSM (OpenStreetMap) is used to crawl map data of the target area, and the data is converted into an initial road network file in net.xm format using a conversion tool adapted to the simulation software. This results in an initial road network that only contains basic topological relationships and whose attribute information is not refined.

[0028] Step S2: Obtain the basic geographic information SHP file of the target area, read the full road network layer information and filter out all ramp candidate data to construct a ramp layer, generate the centerline and actual boundary line of the ramps based on the constructed ramp layer, and further obtain the ramp area, specifically including: Step S2-1: Obtain the basic geographic information SHP file of the target area, read the full road network layer information and filter out all ramp candidate data to construct the ramp layer; Specifically, based on the basic geographic information SHP file of the target area (including basic road network layers such as highway mainlines and interchanges), the full road network layer information of the SHP file is first read using the Python geopandas library, and the core attributes of all linear road segments are extracted, including non-spatial attributes (segment ID, name, road network unit to which it belongs, connecting mainline ID, etc.) and spatial geometric attributes (geometry field).

[0029] Then, based on the unique characteristics of ramps, multi-dimensional combination filtering rules are set to filter out all ramp candidate data from the full road network layer and construct a unique HighwayRamp ramp layer; The filtering rules for obtaining candidate ramp data in step S2-1 include: Rule 1: The topological relationship only connects the main line of the highway and the interchange nodes, with no other non-interchange type connected objects; Rule 2: The geometric shape is a short linear segment with a clear curve or broken line direction. Specifically, there exists a road center point where the angle between the tangents is less than the first threshold angle, regardless of the vector distance.

[0030] Rule 3: The road network unit to which the road segment belongs has a direct coding association with the interchange and the main line of the expressway.

[0031] Step S2-2: Obtain the latitude and longitude coordinate sequence of the road centerline within the ramp layer, and fit it to obtain the ramp centerline; Step S2-3: Based on the preset ramp width parameters, extend symmetrically to both sides with the center line as the reference to generate the original boundary line of the ramp; Step S2-4: Intersect the original boundary line of the ramp with the boundary lines of other roads to determine the actual boundary line of the ramp, and further obtain the ramp area.

[0032] Specifically, the `geometry` field of the ramp is extracted. This field is a LineString type and stores an ordered sequence of latitude and longitude coordinates that constitute the ramp centerline. This coordinate sequence provides the precise geographical location and linear orientation of the entire ramp. Then, the `convertLonLat2XY` function is called to map the actual ramp centerline coordinate sequence to a two-dimensional coordinate system adapted for the simulation environment, resulting in the mapped ramp centerline coordinate sequence. Connecting all points in this mapped coordinate sequence forms the ramp centerline in the simulation environment. Based on preset ramp width parameters, the centerline is extended symmetrically to both sides to generate the ramp's original boundary line. This original boundary line then intersects with the actual road boundary line to further determine the ramp's actual boundary line.

[0033] Step S3: Based on the basic geographic information SHP file, read the lane line layer, generate the first image corresponding to the ramp area, and crop the first image to obtain multiple first sub-images, specifically including: Step S3-1: Read the lane line layer based on the basic geographic information SHP file and generate the first image corresponding to the ramp area; Load raw data containing road lane line geographic information of the target area into GIS software, such as high-precision remote sensing imagery and finely labeled road network data, to generate a first image in TIFF raster format covering high-resolution lane lines. The first image contains associated geographic coordinate information. Step S3-2: Determine the grid size based on the preset ramp width parameters, which can generally be twice the ramp width; Step S3-3: Analyze the first image based on the grid size to obtain multiple candidate sub-images, specifically including: Step S3-3-1: Sample multiple first sampling points are obtained by sampling along the centerline of the ramp; Step S3-3-2: Using the first sampling point as the center and the grid size as the side length, crop the first image to obtain multiple candidate sub-images.

[0034] Step S3-4: Based on the area ratio of the ramp region in all candidate sub-images, select a portion of candidate sub-images as the first sub-image.

[0035] This allows a large TIFF image to be divided into several sub-images according to geographical range, and ensures that the first sub-image can cover the entire width of the ramp. In addition, in some embodiments, the first sampling points need to be selected densely enough to ensure that no complete ramp segment is missed.

[0036] In this way, switching can prevent similar occurrences. Figure 2 The intersection point on the right is larger than the actual lane line.

[0037] Step S4: Cut each first sub-image using a first auxiliary line parallel to the x-axis and / or a second auxiliary line parallel to the y-axis to obtain the number of intersections between each first auxiliary line and / or second auxiliary line and the lane line; In step S4, multiple first auxiliary lines and multiple second auxiliary lines are provided.

[0038] Step S5: Obtain the number of lane lines corresponding to the corresponding first sub-image based on the number of intersections between each first auxiliary line and / or second auxiliary line and the lane line; When using the first auxiliary line and the second auxiliary line for cutting, in step S5, the maximum value of the number of intersections of all auxiliary lines is taken as the number of lane lines.

[0039] Step S6: Based on the number of lane lines in each first sub-image, obtain the number of lanes and lane width within the ramp; like Figure 3 As shown, the SHP file contains a centerline layer for locating roads. This layer stores the vector coordinate sequence of the road centerline through the geometry field, which can accurately anchor the geographical range and direction of the ramp.

[0040] For each sub-image, sampling points are taken at fixed intervals along the centerline of the ramp. The perpendicular line from the centerline at each sampling point is calculated. The perpendicular line is matched with the pixels of the road boundaries on both sides of the ramp, and the geographical distance between the two points is calculated as the road width at that sampling point.

[0041] Extract the pixel coordinates of the start and end points of the ramp centerline in each sub-image and associate them with the corresponding geographic coordinates. Repeat the above steps for all segmented sub-images, summarize the road width data of all sampling points, and take the segment value as the lane width of each segment of the ramp.

[0042] Step S7: Based on the obtained ramp area and the number and width of lanes within the ramp, modify the initial road network file.

[0043] The net.xml file obtained in step S1 contains basic road network topology information, where the attributes of the road segment nodes corresponding to the ramps include link id, type, coordinates of the start and end connection points (x, y), lane information, linear parameters, etc. Using the ramp linear features and lane number / width data obtained in steps S2-S3, the net.xml file is modified to generate a refined ramp simulation road network, as follows: Step S7-1: Modify ramp features, including: Step S7-1-1: The road segment nodes corresponding to the ramps contain linear parameter attributes. Therefore, by searching the ramp IDs marked in the net.xml file, we can locate all the road segment nodes corresponding to the ramps, thereby obtaining the basic attribute data such as the initial start and end connection point coordinates (x, y), length, and curvature of each ramp.

[0044] Step S7-1-2: Using the two-dimensional coordinate sequence of the ramp centerline after coordinate transformation, traverse the initial linear parameters of the corresponding road segment nodes, determine the matching relationship between the initial start and end connection points and the actual ramp connection points, and accurately locate the road segment nodes to be corrected.

[0045] Step S7-1-3: Replace the initial linear attributes of the road segment nodes with the centerline coordinate sequence, segment curvature and precise length of the real ramp, and simultaneously supplement the specific parameters such as ramp slope and superelevation (if any), so as to realize the refined generation of ramp linearity in the simulated road network.

[0046] Step S7-2: Lane information modification, specifically including: Step S7-2-1: Using the obtained starting point, ending point and segmented pixel coordinates of the ramp centerline, connect the segmented road segments, extract the pixel coordinates at the connection points between the ramp and the main line or interchange node, and clarify the adaptation range of the lane parameters.

[0047] Step S7-2-2: The TIF format contains geographic coordinates, converting the pixel coordinates of the road's start and end points into actual geographic coordinates.

[0048] Step S7-2-3: Convert the actual geographic coordinates of the road's starting and ending points into two-dimensional location coordinates of the road's starting and ending points in the simulated road network.

[0049] Step S7-2-4: Nearest Neighbor Matching. Using the converted coordinates of the ramp connection points, locate the junction in the XML file that is closest to the main line or interchange node. Modify the obtained number of lanes, single lane width, and median width into the attributes of the corresponding road segment node and connection junction of the ramp.

[0050] like Figure 3 As shown, by obtaining accurate ramp network information through the entire process, the basic road network XML file can be modified in a targeted manner, and finally, the refined construction of ramp features, lane parameters, and connections with the main line or nodes can be achieved.

[0051] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

Claims

1. A method for generating a simulated road network to achieve fine-grained optimization of ramps, characterized in that, include: Step S1: Obtain the regional OSM road network and perform format conversion to obtain the initial road network file containing the target region; Step S2: Obtain the basic geographic information SHP file of the target area, read the full road network layer information and filter out all ramp candidate data to construct a ramp layer, generate the center line and actual boundary line of the ramp based on the constructed ramp layer, and further obtain the ramp area; Step S3: Based on the basic geographic information SHP file, read the lane line layer, generate the first image corresponding to the ramp area, and crop the first image to obtain multiple first sub-images; Step S4: Cut each first sub-image using a first auxiliary line parallel to the x-axis and / or a second auxiliary line parallel to the y-axis to obtain the number of intersections between each first auxiliary line and / or second auxiliary line and the lane line; Step S5: Obtain the number of lane lines corresponding to the corresponding first sub-image based on the number of intersections between each first auxiliary line and / or second auxiliary line and the lane line; Step S6: Based on the number of lane lines in each first sub-image, obtain the number of lanes and lane width within the ramp; Step S7: Based on the obtained ramp area and the number and width of lanes within the ramp, modify the initial road network file.

2. The method for generating a simulated road network for fine-grained ramp optimization according to claim 1, characterized in that, Step S2 includes: Step S2-1: Obtain the basic geographic information SHP file of the target area, read the full road network layer information and filter out all ramp candidate data to construct the ramp layer; Step S2-2: Obtain the latitude and longitude coordinate sequence of the road centerline within the ramp layer, and fit it to obtain the ramp centerline; Step S2-3: Based on the preset ramp width parameters, extend symmetrically to both sides with the center line as the reference to generate the original boundary line of the ramp; Step S2-4: Intersect the original boundary line of the ramp with the boundary lines of other roads to determine the actual boundary line of the ramp, and further obtain the ramp area.

3. The method for generating a simulated road network for fine-grained ramp optimization according to claim 2, characterized in that, The filtering rules for obtaining candidate ramp data in step S2-1 include: Rule 1: The topological relationship only connects the main line of the highway and the interchange nodes, with no other non-interchange type connected objects; Rule 2: The geometric shape consists of short linear segments with a clear curve or broken line orientation; Rule 3: The road network unit to which the road segment belongs has a direct coding association with the interchange and the main line of the expressway.

4. The method for generating a simulated road network for fine-grained ramp optimization according to claim 3, characterized in that, The specific examples of obvious curves or broken lines in Rule 2 are as follows: There exists a road whose center point has a tangent angle that is less than the first threshold angle, and whose vector distance is less than the first threshold angle.

5. The method for generating a simulated road network for fine-grained ramp optimization according to claim 1, characterized in that, Step S3 includes: Step S3-1: Read the lane line layer based on the basic geographic information SHP file and generate the first image corresponding to the ramp area; Step S3-2: Determine the grid size based on the preset ramp width parameters; Step S3-3: Perform atmospheric treatment on the first image based on the grid size to obtain multiple candidate sub-images; Step S3-4: Based on the area ratio of the ramp region in all candidate sub-images, select a portion of candidate sub-images as the first sub-image.

6. The method for generating a simulated road network for fine-grained ramp optimization according to claim 5, characterized in that, Step S3-3 includes: Step S3-3-1: Sample multiple first sampling points are obtained by sampling along the centerline of the ramp; Step S3-3-2: Using the first sampling point as the center and the grid size as the side length, crop the first image to obtain multiple candidate sub-images.

7. The method for generating a simulated road network for fine-grained ramp optimization according to claim 1, characterized in that, In step S4, multiple first auxiliary lines and multiple second auxiliary lines are provided.

8. The method for generating a simulated road network for fine-grained ramp optimization according to claim 1, characterized in that, When the first auxiliary line and the second auxiliary line are used for cutting, in step S5, the maximum value of the number of intersections of all auxiliary lines is taken as the number of lane lines.

9. A simulated road network generation device for achieving fine-grained optimization of ramps, comprising a memory, a processor, and a program stored in the memory, characterized in that, When the processor executes the program, it implements the method as described in any one of claims 1-8.

10. A storage medium having a program stored thereon, characterized in that, When the program is executed, it implements the method as described in any one of claims 1-8.

Images