Intelligent selection method for tunnel portal light-dark demarcation mileage

Abstract

The application discloses an intelligent selection method for tunnel portal light-dark boundary mileage, relates to the technical field of tunnel construction and design, and comprises the following steps: acquiring a data set of a constructed tunnel portal, wherein the data set comprises topographic point cloud data, route data, tunnel structure size and light-dark boundary mileage; determining characteristic values of the constructed tunnel portal according to a preset characteristic construction strategy based on the data set, and generating positive and negative sample data sets according to the characteristic values and the light-dark boundary mileage; training a machine learning model according to the positive and negative sample data sets to obtain a target machine learning model; uniformly selecting a to-be-designed tunnel mileage along a design center line of a to-be-designed tunnel, determining characteristic values corresponding to the to-be-designed tunnel mileage, inputting the characteristic values corresponding to the to-be-designed tunnel mileage into the target machine learning model, obtaining a classification result sequence, and determining a target light-dark boundary mileage of the to-be-designed tunnel according to the classification result sequence. The application can improve the design efficiency.

Description

Technical Field

[0001] This application relates to the fields of tunnel construction and design technology, and in particular to an intelligent selection method for the boundary mileage between open and closed tunnel entrances. Background Technology

[0002] In tunnel portal design, the selection of the boundary mileage between the open and closed sections directly impacts the construction safety and efficiency of tunnel entry and exit. Traditionally, determining this boundary mileage requires designers to repeatedly calculate the relative positions of the ground and the tunnel based on experience. The main design process involves: 1. Drawing a contour map of the ground; 2. Selecting a mileage along the tunnel's design axis and drawing cross-sectional and longitudinal profiles of the ground elevation; 3. Observing the relative positions of the cross-sectional and longitudinal profiles with the tunnel portal; 4. Repeating steps 2 and 3 until a suitable boundary mileage is found. It is evident that designing the boundary mileage between the open and closed sections of the tunnel portal using traditional methods requires continuous iterative attempts. The entire design process is not only time-consuming and labor-intensive but also highly dependent on the designer's experience.

[0003] The above content is only used to help understand the technical solution of this application and does not represent an admission that the above content is prior art. Summary of the Invention

[0004] The main objective of this application is to provide an intelligent selection method for the boundary mileage between light and dark areas at tunnel entrances, aiming to solve the technical problem that designing the boundary mileage between light and dark areas at tunnel entrances using traditional methods is not only time-consuming and labor-intensive, but also highly dependent on the experience of designers.

[0005] To achieve the above objectives, this application proposes an intelligent selection method for the boundary mileage between light and dark areas at tunnel entrances. The method includes:

[0006] Obtain a dataset of the entrances to constructed tunnels, wherein the dataset includes topographic point cloud data, route data, tunnel structure dimensions, and the mileage of the light-dark boundary;

[0007] Based on the dataset, the feature values ​​of the constructed tunnel entrances are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage.

[0008] A machine learning model is trained based on the positive and negative sample datasets to obtain the target machine learning model;

[0009] The mileage of the tunnel to be designed is uniformly selected along the design centerline of the tunnel to be designed, the feature value corresponding to the mileage of the tunnel to be designed is determined, and the feature value corresponding to the mileage of the tunnel to be designed is input into the target machine learning model to obtain a classification result sequence. The target light-dark boundary mileage of the tunnel to be designed is determined according to the classification result sequence.

[0010] In one embodiment, the step of determining the feature values ​​of the constructed tunnel entrance based on the dataset according to a preset feature construction strategy, and generating a positive and negative sample dataset based on the feature values ​​and the light-dark boundary mileage includes:

[0011] A Delaunay triangulation object is created based on the first horizontal and vertical axis coordinates of the discrete point cloud in the terrain point cloud data.

[0012] Generate the second horizontal and vertical coordinates of the dense point cloud, and determine whether the dense point cloud is inside the Delaunay triangulation object based on the second horizontal and vertical coordinates;

[0013] When a dense point cloud is inside the Delaunay triangulation object, the vertical axis coordinates of the dense point cloud are determined by linear interpolation.

[0014] When the dense point cloud is not inside the Delaunay triangulation object, the nearest discrete point cloud of the dense point cloud is determined according to the second horizontal and vertical axis coordinates, and the first vertical axis coordinate of the nearest discrete point cloud is used as the second vertical axis coordinate of the dense point cloud.

[0015] Determine the terrain density point data based on the second horizontal and vertical axis coordinates and the second vertical axis coordinates;

[0016] A new dataset of tunnel entrances that have been constructed is determined based on the terrain density point data, the route data, the tunnel size data, and the light-dark boundary mileage.

[0017] Based on the new dataset, the feature values ​​of the tunnel entrances that have been constructed are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage.

[0018] In one embodiment, the feature values ​​include a first feature value, a second feature value, a third feature value, and a fourth feature value. The first feature value is the value representing the maximum elevation difference between the longitudinal section topographic line and the tunnel design axis on the longitudinal profile. The second feature value is the elevation difference between the cross-sectional topographic line and the top of the tunnel entrance on the cross-section. The third feature value is the cosine of the inclination angle corresponding to the straight line containing the first and second intersection points on the cross-section. The first intersection point is the intersection of the first straight line and the topographic line, and the first straight line is the straight line obtained by rotating the left bottom vertical line of the tunnel entrance by a preset angle counterclockwise. The second intersection point is the intersection of the second straight line and the topographic line, and the second straight line is the straight line obtained by rotating the right bottom vertical line of the tunnel entrance by a preset angle clockwise. The fourth feature value is the ratio of the tunnel cross-sectional area below the topographic line on the cross-section to the total cross-sectional area of ​​the tunnel. The step of determining the feature values ​​of the constructed tunnel entrance based on the dataset and according to a preset feature construction strategy includes:

[0019] Based on the data of dense terrain points, determine the cross-sectional and longitudinal terrain lines at the entrance of the constructed tunnel;

[0020] The first feature value is determined based on the longitudinal profile topographic lines and the route data;

[0021] The second feature value is determined based on the cross-sectional topographic lines, the route data, and the tunnel structure dimensions;

[0022] The third characteristic value is determined based on the cross-sectional topographic lines and the tunnel structure dimensions.

[0023] The fourth feature value is determined based on the cross-sectional topographic lines and the tunnel structural dimensions.

[0024] In one embodiment, the step of determining the first feature value based on the longitudinal profile topographic line and the route data includes:

[0025] Based on the route data, the tunnel design axis, design start coordinates, and design end coordinates are determined, and multiple discrete points on the tunnel design axis are determined.

[0026] Based on the design start coordinates and the design end coordinates, determine the three-dimensional coordinates of the discrete points on the axis;

[0027] The first feature value is determined based on the three-dimensional coordinates and the longitudinal profile topographic lines.

[0028] In one embodiment, the step of determining the second feature value based on the cross-sectional topographic line, the route data, and the tunnel structure dimensions includes:

[0029] Based on the cross-sectional topographic lines, the route data, and the tunnel structural dimensions, determine the cross-sectional mileage, the design starting mileage, the design starting coordinates, the design ending coordinates, and the tunnel radius.

[0030] Based on the cross-section mileage, design starting mileage, design starting coordinates, and design ending coordinates, the target coordinates are determined, wherein the target coordinates are the coordinates of the centerline of the cross-section tunnel;

[0031] The terrain height value is determined based on the horizontal and vertical axis coordinates of the target coordinates, and the second feature value is determined based on the terrain height value, the vertical axis coordinate of the target coordinates, and the tunnel radius.

[0032] In one embodiment, the step of determining the third feature value based on the cross-sectional topographic line and the tunnel structure dimensions includes:

[0033] Determine the target unit normal vector of the longitudinal section where the tunnel axis is located, and determine the left bottom point and right bottom point of the tunnel entrance based on the target unit normal vector and the tunnel structure dimensions;

[0034] Along the direction of the target unit normal vector, starting from the bottom left point of the tunnel entrance, move left or right by a preset step length to obtain the first current point. Determine the first height difference between the first straight line and the cross-sectional terrain line based on the horizontal and vertical coordinates of the first current point. Within a preset search range, continuously move and update the first current point based on the preset step length until the first height difference determined based on the first current point is less than a preset threshold. Then, determine the vertical coordinate of the first current point based on the horizontal and vertical coordinates of the first current point and the terrain dense point data, and use the horizontal and vertical coordinates of the first current point as the first intersection point.

[0035] Along the direction of the target unit normal vector, with the bottom right point of the tunnel entrance as the second initial point, move left or right by a preset step length to obtain the second current point. Determine the second height difference between the second straight line and the cross-sectional terrain line based on the horizontal and vertical coordinates of the second current point. Within a preset search range, continuously move and update the second current point based on the preset step length until the second height difference determined based on the second current point is less than a preset threshold. Then, determine the vertical coordinate of the second current point based on the horizontal and vertical coordinates of the second current point and the terrain dense point data, and use the horizontal and vertical coordinates of the second current point as the second intersection point.

[0036] The target vector is determined based on the first intersection point and the second intersection point, and the cosine of the angle between the target vector and the target unit normal vector is determined as the third eigenvalue.

[0037] In one embodiment, the step of determining the fourth feature value based on the cross-sectional topographic line and the tunnel structure dimensions includes:

[0038] Starting from the center of the tunnel entrance, the tunnel section is divided into units at preset intervals to the left and right along the radius of the horizontal direction of the tunnel entrance on the cross section.

[0039] Calculate the height difference between the tunnel line and the terrain line in the cross section within each unit segment based on the tunnel structure dimensions.

[0040] The area between the cross-sectional tunnel line and the cross-sectional topographic line is determined based on the height difference, and the area of ​​the overlapping region is determined based on the area.

[0041] The fourth feature value is determined based on the area of ​​the overlapping region and the cross-sectional area of ​​the tunnel.

[0042] In one embodiment, the positive and negative sample dataset includes a positive sample dataset and a negative sample dataset; wherein, the step of generating the positive and negative sample dataset based on the feature values ​​and the light / dark boundary mileage includes:

[0043] The feature values ​​located at the light-dark boundary mileage are identified as positive samples, and the positive sample dataset is constructed based on the positive samples;

[0044] Feature values ​​far from the boundary between light and dark are identified as negative samples, and the negative sample dataset is constructed based on these negative samples.

[0045] In one embodiment, the step of training a machine learning model based on the positive and negative sample dataset to obtain a target machine learning model includes:

[0046] Set the range of values ​​for hyperparameters in the machine learning model;

[0047] On the training set of the positive and negative sample datasets, the optimal hyperparameters are determined from the range of values ​​based on five-fold cross-validation;

[0048] The machine learning model is retrained on the positive and negative sample datasets using the optimal hyperparameters to obtain the target machine learning model.

[0049] In one embodiment, the classification result sequence includes multiple classification results, each including mileage, category, and probability; wherein, the step of determining the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence includes:

[0050] The classification result with a preset value in the classification result sequence is determined as the first classification result;

[0051] The probability ranking number of the first classification result is determined based on the probability of the classification result, and the first classification result whose probability ranking number is less than or equal to the preset ranking value is determined as the second classification result;

[0052] The mileage furthest from the tunnel entrance of the tunnel to be designed in the second classification results is determined as the target light-dark boundary mileage.

[0053] Furthermore, to achieve the above objectives, this application also proposes an intelligent selection method and apparatus for the light-dark boundary mileage of a tunnel entrance, the intelligent selection method and apparatus for the light-dark boundary mileage of a tunnel entrance comprising:

[0054] The acquisition module is used to acquire a dataset of the entrances of constructed tunnels, wherein the dataset includes topographic point cloud data, route data, tunnel structure dimensions, and the mileage of the light-dark boundary.

[0055] The generation module is used to determine the feature values ​​of the constructed tunnel entrances based on the dataset and according to a preset feature construction strategy, and to generate positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage.

[0056] The training module is used to train a machine learning model based on the positive and negative sample dataset to obtain the target machine learning model;

[0057] The determination module is used to uniformly select the mileage of the tunnel to be designed along the design centerline of the tunnel to be designed, determine the feature value corresponding to the mileage of the tunnel to be designed, input the feature value corresponding to the mileage of the tunnel to be designed into the target machine learning model, obtain a classification result sequence, and determine the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence.

[0058] Furthermore, to achieve the above objectives, this application also proposes an intelligent selection method and device for the light-dark boundary mileage of a tunnel entrance. The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor. The computer program is configured to implement the steps of the intelligent selection method for the light-dark boundary mileage of a tunnel entrance as described above.

[0059] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements the steps of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance as described above.

[0060] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance as described above.

[0061] One or more technical solutions proposed in this application have at least the following technical effects:

[0062] The intelligent selection method for the light-dark boundary mileage of tunnel entrances proposed in this application involves acquiring a dataset of constructed tunnel entrances, including topographic point cloud data, route data, tunnel structural dimensions, and light-dark boundary mileage. Based on this dataset, feature values ​​of the constructed tunnel entrances are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage. A machine learning model is trained based on the positive and negative sample datasets to obtain a target machine learning model. The mileage of the tunnel to be designed is uniformly selected along the design centerline of the tunnel to be designed, and the feature values ​​corresponding to the mileage are determined. These feature values ​​are then input into the target machine learning model to obtain a classification result sequence, and the target light-dark boundary mileage of the tunnel to be designed is determined based on the classification result sequence. This invention solves the technical problem that designing the light-dark boundary mileage of tunnel entrances using traditional methods is time-consuming, labor-intensive, and highly dependent on the experience of designers. Compared with existing technologies, this application makes full use of the dataset of already constructed tunnel entrances to extract structural features and uses machine learning to mine the potential relationship between structural features and light-dark boundaries. The design of the tunnel to be designed can be realized without repeated attempts by designers, thus achieving efficient design. Moreover, the design results are not affected by the subjectivity of designers, thereby ensuring the stability and reliability of the design results. Attached Figure Description

[0063] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0064] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 This is a flowchart illustrating an embodiment of the intelligent selection method for the boundary between light and dark areas at the tunnel entrance in this application.

[0066] Figure 2 The interpolation calculation flowchart provided in Embodiment 1 of the intelligent selection method for the boundary mileage between the light and dark sections of the tunnel entrance in this application;

[0067] Figure 3 A schematic diagram of five-fold cross-validation provided for the first embodiment of the intelligent selection method for the boundary mileage between the light and dark areas at the tunnel entrance in this application;

[0068] Figure 4 A technical roadmap provided for an embodiment of the intelligent selection method for the boundary mileage between light and dark tunnel entrances in this application;

[0069] Figure 5 This is a flowchart illustrating the second embodiment of the intelligent selection method for the boundary between light and dark areas at the tunnel entrance in this application.

[0070] Figure 6 This is a first feature illustration provided for Embodiment 2 of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in this application;

[0071] Figure 7 This is a second feature illustration provided in Embodiment 2 of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in this application;

[0072] Figure 8 This is a third feature illustration provided in Embodiment 2 of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in this application;

[0073] Figure 9 This is a schematic diagram illustrating the intersection point calculation principle of the intelligent selection method for the boundary mileage between the light and dark areas at the tunnel entrance in this application, as provided in Embodiment 2.

[0074] Figure 10 A flowchart illustrating the intersection point calculation process provided in Embodiment 2 of the intelligent selection method for the boundary mileage between the light and dark areas at the tunnel entrance in this application;

[0075] Figure 11 This is a diagram illustrating the fourth feature of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in this application, as shown in Embodiment 2.

[0076] Figure 12 This is a topographic contour map of densely constructed tunnel points provided in Embodiment 2 of the intelligent selection method for the boundary mileage between the light and dark sections of the tunnel entrance in this application.

[0077] Figure 13 This is a topographic contour map of densely packed points of the tunnel to be designed, provided in Embodiment 2 of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in this application.

[0078] Figure 14 This is a schematic diagram of the module structure of the intelligent selection device for the light and dark boundary mileage of the tunnel entrance according to an embodiment of this application;

[0079] Figure 15 This is a schematic diagram of the equipment structure of the hardware operating environment involved in the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in the embodiments of this application.

[0080] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0081] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0082] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0083] The main solution of this application embodiment is as follows: First, acquire a dataset of constructed tunnel entrances, wherein the dataset includes topographic point cloud data, route data, tunnel structural dimensions, and the mileage of the light-dark boundary. Second, based on the dataset, determine the feature values ​​of the constructed tunnel entrances according to a preset feature construction strategy, and generate positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage. Third, train a machine learning model based on the positive and negative sample datasets to obtain a target machine learning model. Fourth, uniformly select the mileage of the tunnel to be designed along the design centerline, determine the feature values ​​corresponding to the mileage, input the feature values ​​corresponding to the mileage into the target machine learning model to obtain a classification result sequence, and determine the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence.

[0084] In this embodiment, for ease of description, the following description will focus on the intelligent selection device that identifies the boundary between light and dark areas at the tunnel entrance.

[0085] Because designing the boundary between light and dark areas at tunnel entrances using traditional methods requires continuous iteration and experimentation, the entire design process is not only time-consuming and labor-intensive, but also highly dependent on the experience of the designers.

[0086] This application provides a solution that fully utilizes the dataset of existing tunnel entrances to extract structural features and uses machine learning to mine the potential relationship between structural features and light-dark boundaries. This allows for the design of the tunnel to be designed without repeated attempts by designers, thus achieving efficient design. Furthermore, the design results are not affected by the subjectivity of the designers, thereby ensuring the stability and reliability of the design results.

[0087] As can be seen from the above embodiments, this application obtains a dataset of the entrances of constructed tunnels, wherein the dataset includes topographic point cloud data, line data, tunnel structural dimensions, and the mileage of the light-dark boundary; based on the dataset, the feature values ​​of the entrances of the constructed tunnels are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated according to the feature values ​​and the light-dark boundary mileage; a machine learning model is trained according to the positive and negative sample datasets to obtain a target machine learning model; the mileage of the tunnel to be designed is uniformly selected along the design centerline of the tunnel to be designed, the feature values ​​corresponding to the mileage of the tunnel to be designed are determined, and the feature values ​​corresponding to the mileage of the tunnel to be designed are input into the target machine learning model to obtain a classification result sequence, and the target light-dark boundary mileage of the tunnel to be designed is determined according to the classification result sequence. This invention solves the technical problem that designing the light-dark boundary mileage of tunnel entrances using traditional methods is time-consuming, labor-intensive, and highly dependent on the experience of designers. Compared with existing technologies, this application makes full use of the dataset of already constructed tunnel entrances to extract structural features and uses machine learning to mine the potential relationship between structural features and light-dark boundaries. The design of the tunnel to be designed can be realized without repeated attempts by designers, thus achieving efficient design. Moreover, the design results are not affected by the subjectivity of designers, thereby ensuring the stability and reliability of the design results.

[0088] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device capable of performing the above functions, such as an intelligent selection device for the boundary mileage between light and dark areas at tunnel entrances. The following description uses an intelligent selection device for the boundary mileage between light and dark areas at tunnel entrances as an example to illustrate this embodiment and the subsequent embodiments.

[0089] Based on this, embodiments of this application provide an intelligent selection method for the mileage of the light-dark boundary at the tunnel entrance, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the first embodiment of the intelligent selection method for the boundary mileage between light and dark areas at the tunnel entrance in this application.

[0090] In this embodiment, the intelligent selection method for the tunnel entrance light-dark boundary mileage includes steps S10 to S40:

[0091] Step S10: Obtain the dataset of the tunnel entrances that have been constructed, wherein the dataset includes topographic point cloud data, route data, tunnel structure dimensions, and the mileage of the light-dark boundary.

[0092] It should be noted that the tunnel portal open-cut and cut-cut boundary mileage refers to the boundary mileage between open-cut and cut-cut sections at the tunnel portal. That is, during the excavation process at the tunnel portal, after entering the tunnel portal from the ground, the boundary mileage from open-cut to cut-cut is reached. Structural features can be extracted based on the dataset of already constructed tunnel portals, and machine learning methods can be used to explore their potential relationship with the open-cut and cut-cut boundary mileage, thereby guiding new tunnel designs.

[0093] In one feasible implementation, the step of determining the feature values ​​of the constructed tunnel entrance based on the dataset according to a preset feature construction strategy, and generating positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage, includes: creating a Delaunay triangulation object based on the first horizontal and vertical axis coordinates of discrete point clouds in the terrain point cloud data; generating the second horizontal and vertical axis coordinates of dense point clouds, and determining whether the dense point clouds are inside the Delaunay triangulation object based on the second horizontal and vertical axis coordinates; when the dense point clouds are inside the Delaunay triangulation object, determining the vertical axis coordinates of the dense point clouds through linear interpolation ... When the cloud is not inside the Delaunay triangulation object, the nearest discrete point cloud of the dense point cloud is determined based on the second horizontal and vertical axis coordinates, and the first vertical axis coordinate of the nearest discrete point cloud is used as the second vertical axis coordinate of the dense point cloud; terrain dense point data is determined based on the second horizontal and vertical axis coordinates and the second vertical axis coordinate; a new dataset of the constructed tunnel entrance is determined based on the terrain dense point data, the line data, the tunnel size data, and the light-dark boundary mileage; based on the new dataset, the feature values ​​of the constructed tunnel entrance are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage.

[0094] It should be noted that, due to the sparse and non-uniform nature of the surveyed terrain point cloud data, it needs to be interpolated into more uniform dense terrain point data to obtain the terrain height at any location near the tunnel entrance; the first horizontal and vertical axis coordinates refer to the horizontal and vertical coordinates of the discrete point cloud, and the second horizontal and vertical axis coordinates refer to the horizontal and vertical coordinates of the dense point cloud; triangulation is the process of generating a set of triangles from a given set of planar points. Delaunay triangulation refers to triangulation in which the circumcircles of all triangles satisfy the empty circle property. The empty circle property means that the circumcircle of a triangle does not contain any vertices in the planar point set. The Delaunay triangulation object is the circumcircle of the triangle.

[0095] In specific implementations, such as Figure 2 The interpolation calculation flowchart shown below illustrates the specific calculation steps for generating dense terrain point data based on terrain point cloud data interpolation:

[0096] (1) Create a Delaunay triangulation object based on the surveyed terrain point cloud data, and construct the Delaunay triangulation object according to the x-coordinate (horizontal axis coordinate) and y-coordinate (vertical axis coordinate) of the discrete point cloud in the terrain point cloud data.

[0097] (2) For the x-coordinate (horizontal axis) and y-coordinate (vertical axis) of the generated dense point cloud, if the x-coordinate and y-coordinate of the dense point cloud are inside the Delaunay triangulation object, the z-coordinate (vertical axis) of the dense point cloud is directly calculated by linear interpolation. Otherwise, the z-coordinate (vertical axis) of the nearest neighbor of the dense point cloud is used as the z-coordinate (vertical axis) of the dense point cloud. The nearest neighbor is the neighboring point cloud in the terrain point cloud data that is closest to the dense point cloud. The specific calculation method of the nearest neighbor is as follows:

[0098] min{(x1-x0) 2 +(y1-y0) 2 (x2-x0) 2 +(y2-y0) 2 ,...,(x i -x0) 2 +(y i -y0) 2}

[0099] In the formula, i is the data number of the discrete point cloud, (x i ,y i (x0, y0) represents the x and y coordinates of the i-th point, and (x0, y0) represents the x and y coordinates of the dense point cloud.

[0100] In this embodiment, when generating dense terrain point data based on discrete terrain point cloud data, the discrete point cloud closest to the dense point cloud is selected for interpolation to repair the hollow triangles in the Delaunay triangulation, thereby achieving the effect of interpolation for any point.

[0101] Step S20: Based on the dataset, determine the feature values ​​of the constructed tunnel entrance according to a preset feature construction strategy, and generate positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage.

[0102] It should be noted that, according to the preset feature construction strategy, the characteristic values ​​of the cross section, longitudinal section and the relative position of the tunnel at the entrance of the constructed tunnel can be determined. These mainly include four characteristic values: the first characteristic value is the maximum difference between the terrain height and the tunnel design axis on the longitudinal section; the second characteristic value is the difference between the terrain height and the top of the tunnel entrance on the cross section; the third characteristic value is the cosine value of the inclination angle corresponding to the line where the intersection of the straight line at the bottom of the tunnel entrance at 45° to the left and right and the terrain line is located on the cross section; and the fourth characteristic value is the ratio of the tunnel cross-sectional area below the terrain line to the total cross-sectional area of ​​the tunnel.

[0103] In one feasible implementation, the positive and negative sample dataset includes a positive sample dataset and a negative sample dataset; wherein, the step of generating the positive and negative sample dataset based on the feature values ​​and the light-dark boundary mileage includes: determining the feature values ​​located at the light-dark boundary mileage as positive samples, and constructing the positive sample dataset based on the positive samples; determining the feature values ​​far from the light-dark boundary as negative samples, and constructing the negative sample dataset based on the negative samples.

[0104] It should be noted that both positive and negative samples include feature values ​​and sample codes.

[0105] In the specific implementation, when constructing the positive sample dataset, the mileage of the light-dark boundary at the tunnel entrance that has been constructed is selected, and four feature values ​​at the light-dark boundary at the tunnel entrance are determined. This sample is then encoded as 1 as a positive sample.

[0106] In the specific implementation, when constructing the negative sample dataset, locations (greater than 20m) far from the light-dark boundary of the tunnel entrance are selected discretely. Four feature values ​​of these locations are calculated, and the sample is encoded as 0 as a negative sample.

[0107] In this embodiment, by extracting positive samples from the light-dark boundary mileage of the tunnel entrance and negative samples from the light-dark boundary mileage far from the tunnel entrance, not only can more high-quality datasets be constructed, but the accuracy and robustness of the model can also be improved based on a larger amount of training data.

[0108] Step S30: Train a machine learning model based on the positive and negative sample dataset to obtain the target machine learning model;

[0109] It should be noted that after obtaining the positive and negative sample datasets, 80% of the positive and negative sample datasets can be randomly divided into the training set, and the remaining 20% ​​can be divided into the test set. Then, a machine learning model can be built based on the training set. In order to obtain the optimal model, the optimal hyperparameters of the machine learning model can be adjusted based on grid search and five-fold cross-validation, and then the target machine learning model can be built based on the training set.

[0110] In one feasible implementation, the step of training a machine learning model based on the positive and negative sample datasets to obtain a target machine learning model includes: setting the range of values ​​for hyperparameters in the machine learning model; determining the optimal hyperparameters from the range of values ​​on the training set of the positive and negative sample datasets based on five-fold cross-validation; and retraining the machine learning model on the positive and negative sample datasets using the optimal hyperparameters to obtain the target machine learning model.

[0111] In practice, the specific implementation steps of the target machine learning model are as follows:

[0112] (1) Split the dataset. Merge the positive sample dataset and the negative sample dataset, shuffle their order, and randomly select 80% of the dataset as the training set and 20% as the test set.

[0113] (2) Select a machine learning algorithm to generate a machine learning model. The machine learning algorithms that can be selected include, but are not limited to, the following algorithms: decision tree, SVM, K-nearest neighbors, random forest, Adaboost, GBDT, XGBoost, CatBoost, LightGBM, BP neural network, etc.

[0114] (3) Finding the optimal hyperparameters for the machine learning model:

[0115] The range of values ​​for hyperparameters in the machine learning model is set, and the optimal hyperparameters are searched based on grid search and five-fold cross-validation. The process of five-fold cross-validation is as follows: Figure 3 As shown, the positive and negative sample datasets are divided into five subsets. In each iteration, one subset is selected as the test set without repetition, and the other four subsets are used as the training set to build a machine learning model. The accuracy of the machine learning model on the test set is taken as the model accuracy of that round, and the average accuracy of the model over five rounds is taken as the model accuracy of that set of hyperparameters.

[0116] (4) Establish the target machine learning model based on the optimal hyperparameters:

[0117] The hyperparameters that have the highest model accuracy are determined as the optimal hyperparameters, and the target machine learning model (i.e. the optimal machine learning classification model) is re-engineered on the entire training set using the optimal hyperparameters.

[0118] Step S40: Uniformly select the mileage of the tunnel to be designed along the design center line of the tunnel to be designed, determine the feature value corresponding to the mileage of the tunnel to be designed, input the feature value corresponding to the mileage of the tunnel to be designed into the target machine learning model to obtain a classification result sequence, and determine the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence.

[0119] In the specific implementation, for the tunnel to be designed, the mileage can be selected evenly and discretely along the design centerline. For example, the mileage positions can be selected evenly at a preset interval along the design centerline. Preferably, the preset interval can be set to 1m. After selecting the mileage of the tunnel to be designed, the feature value at each mileage is calculated. For example, the feature value corresponding to the selected mileage is determined according to a preset feature construction strategy. Finally, the feature value corresponding to the mileage is input into the trained machine learning model (i.e., the target machine learning model) to obtain the classification result sequence. Finally, the target light-dark boundary mileage is determined based on the classification result sequence. The classification result sequence includes multiple classification results, and each classification result includes data such as mileage, category, and probability.

[0120] In specific implementations, such as Figure 4 The technical roadmap shown illustrates the following steps for obtaining the target light-dark boundary mileage of the tunnel to be designed:

[0121] (1) Collect the dataset of the tunnel entrances to be designed;

[0122] (2) Based on point cloud data interpolation calculation, generate terrain dense point data of terrain point cloud data;

[0123] (3) Discretely extract feature values ​​at different mileages along the design axis and input them into the trained machine learning model to obtain a classification result sequence;

[0124] (4) Search the classification result sequence and obtain the mileage of the light and dark boundary at the tunnel entrance.

[0125] In one feasible implementation, the classification result sequence includes multiple classification results, each including mileage, category, and probability; wherein, the step of determining the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence includes: determining the classification result with a preset value in the classification result sequence as the first classification result; determining the probability sorting number of the first classification result based on the probability of the classification result, and determining the first classification result with a probability sorting number less than or equal to the preset sorting value as the second classification result; and determining the mileage of the second classification result that is farthest from the tunnel entrance of the tunnel to be designed as the target light-dark boundary mileage.

[0126] It should be noted that the categories in the classification results include category 1 and category 0; the preset value is 1, which means that the classification result with category 1 in the classification result sequence is determined as the first classification result; the higher the probability of the first classification result, the higher the probability ranking number of the first classification result, that is, the probability ranking number of the first classification result is determined in descending order of probability, for example, the probability ranking number of the first classification result with the highest probability is 1, and the probability ranking number of the second highest probability first classification result is 2; the preset ranking value is 5.

[0127] In the specific implementation, the classification result with category 1 in the classification result sequence is determined as the first classification result. The top five classification results with the highest probability in the first classification result are retained as the second classification result. The mileage corresponding to the second classification result is the potential tunnel entrance light-dark boundary mileage. Then, the mileage farthest from the tunnel entrance mileage is selected from the potential tunnel entrance light-dark boundary mileage as the target light-dark boundary mileage. The selection rule for the mileage farthest from the tunnel entrance is as follows: if the light-dark boundary mileage of the tunnel entrance is to be designed, the mileage corresponding to the classification result with the largest mileage position is selected; if the light-dark boundary mileage of the tunnel exit is to be designed, the mileage corresponding to the classification result with the smallest mileage position is selected.

[0128] In this embodiment, a dataset of existing tunnel entrances is acquired, including topographic point cloud data, route data, tunnel structural dimensions, and the mileage of the light-dark boundary. Based on the dataset, feature values ​​of the existing tunnel entrances are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage. A machine learning model is trained based on the positive and negative sample datasets to obtain a target machine learning model. The mileage of the tunnel to be designed is uniformly selected along the design centerline of the tunnel to be designed, and the feature values ​​corresponding to the mileage of the tunnel to be designed are determined. The feature values ​​corresponding to the mileage of the tunnel to be designed are input into the target machine learning model to obtain a classification result sequence, and the target light-dark boundary mileage of the tunnel to be designed is determined based on the classification result sequence. This invention solves the technical problem that designing the light-dark boundary mileage of tunnel entrances using traditional methods is time-consuming, labor-intensive, and highly dependent on the experience of designers. Compared with existing technologies, this application makes full use of the dataset of already constructed tunnel entrances to extract structural features and uses machine learning to mine the potential relationship between structural features and light-dark boundaries. The design of the tunnel to be designed can be realized without repeated attempts by designers, thus achieving efficient design. Moreover, the design results are not affected by the subjectivity of designers, thereby ensuring the stability and reliability of the design results.

[0129] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment can be referred to the above description, and will not be repeated hereafter. Based on this, the feature value includes a first feature value, a second feature value, a third feature value, and a fourth feature value. The first feature value is the maximum height difference between the longitudinal section topographic line and the tunnel design axis on the longitudinal section. The second feature value is the height difference between the cross-sectional topographic line and the top of the tunnel entrance on the cross section. The third feature value is the cosine of the inclination angle corresponding to the straight line where the first and second intersection points are located on the cross section. The first intersection point is the intersection of the first straight line and the topographic line. The first straight line is the straight line after rotating the left bottom vertical line of the tunnel entrance counterclockwise by a preset angle. The second intersection point is the intersection of the second straight line and the topographic line. The second straight line is the straight line after rotating the right bottom vertical line of the tunnel entrance clockwise by a preset angle. The fourth feature value is the ratio of the tunnel cross-sectional area below the topographic line on the cross section to the total tunnel cross-sectional area. Please refer to [reference needed]. Figure 5 Step S20 includes steps S201 to S205:

[0130] Step S201: Determine the cross-sectional and longitudinal topographic lines at the entrance of the constructed tunnel based on the topographic density data.

[0131] Step S202: Determine the first feature value based on the longitudinal profile topographic line and the route data;

[0132] In one feasible implementation, the step of determining the first feature value based on the longitudinal profile topographic line and the route data includes: determining the tunnel design axis, design start coordinates, and design end coordinates based on the route data, and determining multiple discrete points on the tunnel design axis; determining the three-dimensional coordinates of the discrete points based on the design start coordinates and the design end coordinates; and determining the first feature value based on the three-dimensional coordinates and the longitudinal profile topographic line.

[0133] In specific implementations, such as Figure 6 As shown, the first characteristic value is the value with the largest difference between the terrain height and the tunnel design axis on the longitudinal section (that is, the value with the largest difference between the terrain line height and the tunnel design axis on the longitudinal section). The specific calculation steps for the first characteristic value are as follows:

[0134] (1) For the tunnel design axis, the spatial straight line segment between the design start point coordinates and the design end point coordinates is discretized according to a preset discretization interval to obtain multiple axis discrete points on the tunnel design axis, and the three-dimensional coordinates of the axis discrete points (i.e., the horizontal, vertical and horizontal axis coordinates of the axis discrete points) are determined. Preferably, the preset discretization interval can be set to 0.1m. Specifically, the specific calculation method of the three-dimensional coordinates is as follows:

[0135] p j=p s +j·0.1·(p e -p s ) / |p e -p s |

[0136] In the formula, p j Let p be the coordinates of the j-th discrete point along the tunnel design axis. e To design the starting coordinates, p s To design the termination coordinates, |p e -p s | is vector p e -p s The model, p e p s p j Each point is a vector in the form of (x, y, z).

[0137] (2) Then, calculate the maximum elevation difference between the longitudinal profile topographic line and the tunnel design axis (i.e., the first characteristic value) based on the discrete points of the axis. The specific calculation method is as follows:

[0138] f1 = max{g 1z -p 1z ,g 2z -p 2z ,...,g nz -p nz}

[0139] In the formula, p nz Let the z-coordinate and g-coordinate of the nth discrete point on the tunnel axis be defined. nz To determine the z-coordinate value of the topographic point on the longitudinal profile topographic line based on the horizontal and vertical coordinates of the nth discrete point on the axis, f1 is the first characteristic value.

[0140] Step S203: Determine the second feature value based on the cross-sectional topographic line, the route data, and the tunnel structure dimensions;

[0141] In one feasible implementation, the step of determining the second feature value based on the cross-sectional topographic line, the route data, and the tunnel structural dimensions includes: determining the cross-sectional mileage, design starting mileage, design starting coordinates, design ending coordinates, and tunnel radius based on the cross-sectional topographic line, the route data, and the tunnel structural dimensions; determining the target coordinates based on the cross-sectional mileage, the design starting mileage, the design starting coordinates, and the design ending coordinates, wherein the target coordinates are the coordinates of the centerline of the cross-sectional tunnel; determining the terrain height value based on the horizontal and vertical axis coordinates of the target coordinates; and determining the second feature value based on the terrain height value, the vertical axis coordinate of the target coordinates, and the tunnel radius.

[0142] In specific implementations, such as Figure 7 As shown, the second characteristic value is the difference between the terrain height on the cross section and the top of the tunnel entrance (i.e., the difference between the terrain line on the cross section and the top of the tunnel entrance). The specific calculation steps for the second characteristic value are as follows.

[0143] (1) First, determine the cross-sectional mileage, design starting mileage, design starting coordinates, design ending coordinates, and tunnel radius by using the cross-sectional topographic lines, route data, and tunnel structural dimensions. Then, determine the coordinates of the tunnel centerline (i.e., the target coordinates) based on the cross-sectional mileage, design starting mileage, design starting coordinates, design ending coordinates, and tunnel radius. The specific calculation steps are as follows:

[0144] p c =p s +(m c -m s )·(p e -p s ) / |p e -p s |

[0145] In the formula, m c For cross-sectional mileage, m s To design the starting mileage, p c p represents the coordinates of the tunnel centerline in the cross section. e To design the starting coordinates, p s To design the termination coordinates.

[0146] (2) Then, the second feature value is determined based on the vertical axis coordinate of the target coordinates, the terrain height value, and the tunnel radius. The specific calculation steps are as follows:

[0147] f2 = g cz -p cz -R

[0148] In the formula, p cz The z-coordinate of the tunnel centerline point in the cross section (i.e., the vertical axis coordinate of the target coordinate), g cz R is the terrain height value (i.e., the vertical coordinate of the terrain point on the cross section determined by the horizontal and vertical coordinates of the target coordinates), R is the tunnel radius, and f2 is the second characteristic value.

[0149] Step S204: Determine the third feature value based on the cross-sectional topographic line and the tunnel structure dimensions;

[0150] It should be noted that, as Figure 8As shown, the third characteristic value is the cosine of the inclination angle corresponding to the line where the intersection of the straight line at the bottom of the tunnel at a 45° angle with the topographic line is located. The third characteristic value can well reflect the terrain slope. The key to calculating the third characteristic value is to calculate the first intersection point and the second intersection point (the first intersection point is point A in the figure, and the second intersection point is point B in the figure), and then determine the third characteristic value based on the first intersection point and the second intersection point.

[0151] In one feasible implementation, the step of determining the third feature value based on the cross-sectional topographic line and the tunnel structure dimensions includes: determining the target unit normal vector of the longitudinal section where the tunnel axis is located, and determining the left bottom point and the right bottom point of the tunnel entrance based on the target unit normal vector and the tunnel structure dimensions; moving left or right by a preset step length along the direction of the target unit normal vector, with the left bottom point of the tunnel entrance as the first initial point, to obtain the first current point; determining the first height difference between the first straight line and the cross-sectional topographic line based on the horizontal and vertical coordinates of the first current point; continuously moving and updating the first current point based on the preset step length within a preset search range until the first height difference determined based on the first current point is less than a preset threshold; determining the vertical coordinate of the first current point based on the horizontal and vertical coordinates of the first current point and the topographic dense point data; and then... The horizontal and vertical coordinates of the first current point are used as the first intersection point. Following the direction of the target unit normal vector, the second current point is obtained by moving left or right by a preset step length, starting from the bottom right point of the tunnel entrance. The second height difference between the second straight line and the cross-sectional terrain line is determined based on the horizontal and vertical coordinates of the second current point. Within a preset search range, the second current point is continuously moved and updated based on the preset step length until the second height difference determined by the second current point is less than a preset threshold. Then, the vertical coordinates of the second current point are determined based on the horizontal and vertical coordinates of the second current point and the terrain dense point data. The horizontal and vertical coordinates of the second current point are used as the second intersection point. The target vector is determined based on the first and second intersection points, and the cosine of the angle between the target vector and the target unit normal vector is determined as the third feature value.

[0152] Understandably, since terrain elevation varies irregularly, directly determining the first and second intersection points by fitting the cross-sectional terrain has a large error and is not very universal (different terrains require different fitting functions). Therefore, the first and second intersection points can be determined based on numerical solutions, and then the third eigenvalue can be determined based on the first and second intersection points.

[0153] In the specific implementation, the calculation methods for the first and second intersection points are as follows:

[0154] (1) Calculate the unit normal vector (i.e. the target unit normal vector) of the longitudinal section where the tunnel design axis is located. Since the longitudinal section is a vertical plane, one vector on the plane can be set as (0,0,-1), and another vector on the plane can be obtained from the difference between the design end coordinate and the design start coordinate. The unit normal vector of the longitudinal section can be obtained by cross product and normalization of the two vectors.

[0155] (2) Calculate the coordinates of the left bottom point of the tunnel entrance and the coordinates of the right bottom point of the tunnel entrance, such as Figure 8 As shown, Figure 8 L at the bottom of the middle tunnel is the bottom left point of the tunnel entrance. Figure 8 The bottom R of the middle tunnel is the bottom point Y of the tunnel entrance, and the specific calculation method is as follows:

[0156] p L =p o -d1·p n -(0,0,d2)

[0157] In the formula, p L Let p be the coordinates of the bottom left point of the tunnel entrance. n Let d1 be the target unit normal vector, d2 be the distance from the center of the tunnel to the edge, and d2 be the distance from the center of the tunnel to the bottom.

[0158] p R =p o +d1·p n -(0,0,d2)

[0159] In the formula, p R The coordinates are the coordinates of the bottom right point of the tunnel entrance.

[0160] (3) Calculate the coordinates of the first intersection point based on numerical solution: such as Figure 9 As shown, the main idea is to follow the longitudinal section normal vector p n The direction of the target unit normal vector is moved to the left and right by a small step (step = 0.02) from the bottom point (left bottom point of the tunnel entrance) to obtain the first current point, and the first current point is used as the potential intersection point (p = None). The height difference between the cross-sectional terrain line and the 45° angle line (the first straight line) is calculated using the horizontal and vertical coordinates of the first current point, and this height difference is regarded as the distance error e (first height difference) of the potential intersection point. Then, the first current point is continuously moved within the calculation range S (i.e., the preset search range) until the distance error of the first current point is close to zero (or less than the preset threshold, the preset threshold is 0.01). Then, the horizontal and vertical coordinates of the first current point with the first height difference less than the preset threshold are used as the horizontal and vertical coordinates of the first intersection point, and the z coordinate value of the first intersection point (i.e., the vertical coordinate of the first intersection point) is obtained based on the terrain dense point data; such as Figure 10The flowchart shown above illustrates the calculation process for the first intersection point. Parameter initialization can begin, such as step size step = 0.02, loop count k = 0, maximum search range S = 100, minimum distance deviation e = inf, and potential intersection point p = None. If the current point's position is within the maximum search range, continue moving the current point to the left or right until the distance error of the current point is less than the minimum distance deviation value, then output the current point. Otherwise, continue moving the current point after updating its position and deviation.

[0161] (4) Calculate the coordinates of the second intersection point based on numerical solution (the calculation method of the second intersection point is similar to that of the first intersection point): Calculate the coordinates of the second intersection point based on numerical solution: The main idea is to calculate the coordinates of the second intersection point along the longitudinal section normal vector p. n The direction of the target unit normal vector is moved to the left and right by a small step (step = 0.02) from the bottom point (right bottom point of the tunnel entrance) to obtain the second current point. The second current point is used as the potential intersection point. The height difference between the cross-sectional terrain line and the 45° angle line (the second line) is calculated using the horizontal and vertical coordinates of the second current point. This height difference is regarded as the distance error e (second height difference) of the potential intersection point. Then, the second current point is continuously moved within the calculation range S (i.e., the preset search range) until the distance error of the second current point is close to zero (or less than the preset threshold, which is 0.01). When the second height difference is less than the preset threshold, the horizontal and vertical coordinates of the second current point are used as the horizontal and vertical coordinates of the second intersection point. The z-coordinate value of the second intersection point (i.e., the vertical coordinate of the second intersection point) is obtained based on the terrain dense point data.

[0162] (5) Calculate the target vector formed by the first intersection point and the second intersection point, and determine the cosine value of the angle between the target vector and the longitudinal section unit normal vector (target unit normal vector). The specific calculation method is as follows:

[0163]

[0164] In the formula, p A p represents the first intersection point. B f3 represents the second intersection point, and f3 is the third eigenvalue.

[0165] Step S205: Determine the fourth feature value based on the cross-sectional topographic line and the tunnel structure dimensions.

[0166] In one feasible implementation, the step of determining the fourth characteristic value based on the cross-sectional topographic line and the tunnel structural dimensions includes: taking the center of the tunnel entrance as the starting point, and dividing the tunnel cross-section into unit segments at preset intervals along the radius of the horizontal direction of the tunnel entrance on the cross-section; calculating the height difference between the cross-sectional tunnel line and the cross-sectional topographic line within each unit segment based on the tunnel structural dimensions; determining the area between the cross-sectional tunnel line and the cross-sectional topographic line based on the height difference, and determining the overlapping area based on the area; and determining the fourth characteristic value based on the overlapping area and the tunnel cross-sectional area.

[0167] In specific implementations, such as Figure 11 As shown, the fourth characteristic is the ratio of the tunnel cross-sectional area below the cross-sectional topographic line to the total tunnel cross-sectional area. The specific calculation steps for the fourth characteristic value are as follows:

[0168] (1) Starting from the center of the tunnel, divide the tunnel cross-section into unit segments with a radius of 0.01m (i.e., a preset interval) to the left and right along the horizontal direction of the tunnel entrance on the cross-section. Calculate the height difference between the tunnel line and the terrain line on the cross-section within each unit segment. The specific calculation method is as follows:

[0169]

[0170] In the formula, i is the unit segment number, g iz g is the height of the cross-sectional topographic line in this unit segment. oz This is the height of the center point of the cross-section tunnel.

[0171] (2) Based on the elevation difference, the area between the cross-sectional topographic line and the cross-sectional tunnel line in each unit segment can be calculated to obtain the area of ​​the region between the cross-sectional topographic line and the cross-sectional tunnel line. The overlapping area can then be obtained. The ratio of the overlapping area to the tunnel cross-sectional area is the value of feature four. The specific calculation method is as follows:

[0172]

[0173] In the formula, l is the number of unit segments on the left, r is the number of unit segments on the right, and f4 is the fourth eigenvalue.

[0174] This embodiment determines the cross-sectional and longitudinal topographic lines at the entrance of an existing tunnel based on the topographic density point data; determines the first feature value based on the longitudinal topographic line and the route data; determines the second feature value based on the cross-sectional topographic line, the route data, and the tunnel structural dimensions; determines the third feature value based on the cross-sectional topographic line and the tunnel structural dimensions; and determines the fourth feature value based on the cross-sectional topographic line and the tunnel structural dimensions. Through this method, this embodiment extracts multiple structural features related to the light-dark boundary mileage based on the existing tunnel entrance dataset, and explores the potential relationship between structural features and the light-dark boundary mileage, thereby guiding new tunnel designs.

[0175] The first and second embodiments are described in detail below using specific tunnel entrance design datasets:

[0176] (1) Collect the dataset of tunnel entrances that have been constructed (taking three tunnel entrance datasets as an example), which includes the topographic point cloud data of the tunnel entrances in a text file in the format of ".xyz";

[0177]

[0178] (2) Generate dense terrain point data based on interpolation calculation of terrain point cloud data, such as Figure 12 The topographic contour lines of the topographic dense point data of the three completed tunnel portal datasets are shown below;

[0179] (3) For the three existing tunnel portal datasets, multiple mileages were selected from the three existing tunnels, and four structural feature values ​​of the mileages were extracted. Then, the feature values ​​at the light-dark boundary mileages were encoded as 1 (i.e., positive samples). At intervals of 20m, some mileages far from the light-dark boundary mileages were selected, and their corresponding feature values ​​were encoded as 0 (i.e., negative samples). The established positive and negative sample datasets are shown in Table 1.

[0180] Table 1 Positive and Negative Sample Dataset

[0181]

[0182]

[0183] (4) The positive and negative sample datasets were shuffled, and 80% of the data was randomly selected as the training set and 20% as the test set. A machine learning model was built based on the random forest algorithm. The hyperparameters of the model were adjusted in the algorithm as shown in Table 2. The optimal hyperparameters of the model were obtained on the training set based on five-fold cross-validation. The optimal hyperparameters were 100, 7, and 3. Finally, the optimal hyperparameters were used to build a random forest model on the training set. The evaluation results of the model on the training set and validation set are shown in Table 3. The accuracy of the model on the training set and the test set exceeded 95% in all three aspects, which shows that the established light and dark boundary mileage prediction model is good.

[0184] Table 2. Adjusted Hyperparameters

[0185] Number of trees Maximum depth of tree Minimum number of samples in a leaf node 10,50,100,150,200 1,3,5,7,9 3,5,7,10

[0186] Table 3 Model Evaluation Results

[0187] Dataset accuracy Accuracy Recall rate training set 97.64％ 97.43％ 97.64％ test set 95.77％ 95.66％ 95.77％

[0188] (5) For the tunnel to be designed, dense terrain point data is generated based on interpolation calculation of terrain point cloud data, and the corresponding terrain contour lines are as follows: Figure 13 As shown.

[0189] (6) Select mileage at 1m intervals along the design axis of the tunnel to be designed, calculate the corresponding four feature values, and input the four feature values ​​into the trained machine learning model. The output classification result sequence is shown in Table 4.

[0190] Table 4 Prediction Results

[0191]

[0192]

[0193]

[0194] (7) Based on the classification result sequence, first extract the classification result with predicted category 1. The mileage range corresponding to the classification result is DK321+151~DK321+159. Further select the mileages corresponding to the five elements with the highest positive probability in this range as DK321+153~DK321+157. Considering that the tunnel entrance light and dark boundary mileage needs to be designed, the largest mileage, DK321+157, should be selected. This mileage is the intelligently designed tunnel entrance light and dark boundary mileage.

[0195] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the intelligent selection method for the boundary mileage of the tunnel entrance. Any simple modifications based on this technical concept are within the protection scope of this application.

[0196] This application also provides an intelligent selection device for the boundary mileage between light and dark areas at tunnel entrances. Please refer to [link / reference]. Figure 14 The intelligent selection device for determining the boundary between light and dark areas at the tunnel entrance includes:

[0197] The acquisition module 10 is used to acquire a dataset of the entrance to the constructed tunnel, wherein the dataset includes topographic point cloud data, route data, tunnel structure dimensions, and the mileage of the light-dark boundary.

[0198] The generation module 20 is used to determine the feature values ​​of the constructed tunnel entrance based on the dataset and according to a preset feature construction strategy, and to generate positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage.

[0199] Training module 30 is used to train a machine learning model based on the positive and negative sample dataset to obtain the target machine learning model;

[0200] The determination module 40 is used to uniformly select the mileage of the tunnel to be designed along the design center line of the tunnel to be designed, determine the feature value corresponding to the mileage of the tunnel to be designed, input the feature value corresponding to the mileage of the tunnel to be designed into the target machine learning model, obtain a classification result sequence, and determine the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence.

[0201] The intelligent selection device for determining the boundary mileage of tunnel entrances, provided in this application, employs the intelligent selection method for determining the boundary mileage of tunnel entrances described in the above embodiments. This method solves the technical problem that designing the boundary mileage of tunnel entrances using traditional methods is time-consuming, labor-intensive, and highly dependent on the experience of designers. Compared with the prior art, the beneficial effects of the intelligent selection device for determining the boundary mileage of tunnel entrances provided in this application are the same as those of the intelligent selection method for determining the boundary mileage of tunnel entrances provided in the above embodiments. Furthermore, other technical features of the intelligent selection device for determining the boundary mileage of tunnel entrances are the same as those disclosed in the methods of the above embodiments, and will not be elaborated upon here.

[0202] This application provides an intelligent selection device for the light-dark boundary mileage of a tunnel entrance. The intelligent selection device for the light-dark boundary mileage of a tunnel entrance includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the intelligent selection method for the light-dark boundary mileage of the tunnel entrance in the first embodiment described above.

[0203] The following is for reference. Figure 15The diagram illustrates a structural schematic of an intelligent selection device suitable for implementing the tunnel entrance light-dark boundary mileage in the embodiments of this application. The intelligent selection device for tunnel entrance light-dark boundary mileage in the embodiments of this application may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital radio receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers. Figure 15 The intelligent selection device for the light and dark boundary mileage of the tunnel entrance shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.

[0204] like Figure 15 As shown, the intelligent selection device for the tunnel entrance light-dark boundary mileage may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the intelligent selection device for the tunnel entrance light-dark boundary mileage. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the intelligent selection device for the tunnel portal light / dark boundary mileage to communicate wirelessly or wiredly with other devices to exchange data. Although the figure shows an intelligent selection device for the tunnel portal light / dark boundary mileage with various systems, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems can be implemented alternatively.

[0205] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.

[0206] The intelligent selection device for the tunnel entrance light-dark boundary mileage provided in this application adopts the intelligent selection method for the tunnel entrance light-dark boundary mileage in the above embodiments, which can solve the technical problem of intelligent selection of the tunnel entrance light-dark boundary mileage. Compared with the prior art, the beneficial effects of the intelligent selection device for the tunnel entrance light-dark boundary mileage provided in this application are the same as the beneficial effects of the intelligent selection method for the tunnel entrance light-dark boundary mileage provided in the above embodiments, and other technical features in the intelligent selection device for the tunnel entrance light-dark boundary mileage are the same as the features disclosed in the previous embodiment method, and will not be repeated here.

[0207] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.

[0208] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0209] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the intelligent selection method for the light and dark boundary mileage of the tunnel entrance in the above embodiments.

[0210] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.

[0211] The aforementioned computer-readable storage medium may be included in the intelligent selection device for the light-dark boundary mileage of the tunnel entrance; or it may exist independently and not be assembled into the intelligent selection device for the light-dark boundary mileage of the tunnel entrance.

[0212] The aforementioned computer-readable storage medium carries one or more programs. When these programs are executed by an intelligent selection device for the light-dark boundary mileage of a tunnel entrance, the intelligent selection device for the light-dark boundary mileage of a tunnel entrance performs the following actions: acquires a dataset of constructed tunnel entrances, wherein the dataset includes topographic point cloud data, line data, tunnel structural dimensions, and light-dark boundary mileage; based on the dataset, determines the feature values ​​of the constructed tunnel entrances according to a preset feature construction strategy, and generates positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage; trains a machine learning model based on the positive and negative sample datasets to obtain a target machine learning model; uniformly selects the mileage of the tunnel to be designed along the design centerline of the tunnel to be designed, determines the feature values ​​corresponding to the mileage of the tunnel to be designed, inputs the feature values ​​corresponding to the mileage of the tunnel to be designed into the target machine learning model to obtain a classification result sequence, and determines the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence.

[0213] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0214] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0215] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0216] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the intelligent selection method for the light-dark boundary mileage of the tunnel entrance described above. This solves the technical problem that designing the light-dark boundary mileage of tunnel entrances using traditional methods is time-consuming, labor-intensive, and highly dependent on the experience of designers. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as those of the intelligent selection method for the light-dark boundary mileage of tunnel entrances provided in the above embodiments, and will not be repeated here.

[0217] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the intelligent selection method for the light and dark boundary mileage of the tunnel entrance as described above.

[0218] The computer program product provided in this application can solve the technical problem that designing the light and dark boundary mileage of tunnel entrances using traditional methods is time-consuming, labor-intensive, and highly dependent on the experience of designers. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the intelligent selection method for the light and dark boundary mileage of tunnel entrances provided in the above embodiments, and will not be repeated here.

[0219] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for intelligently selecting the mileage of the light-dark boundary at a tunnel entrance, characterized in that, The method includes: Obtain a dataset of the entrances to constructed tunnels, wherein the dataset includes topographic point cloud data, route data, tunnel structure dimensions, and the mileage of the light-dark boundary; Based on the dataset, the feature values ​​of the constructed tunnel entrances are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage. A machine learning model is trained based on the positive and negative sample datasets to obtain the target machine learning model; The mileage of the tunnel to be designed is uniformly selected along the design centerline of the tunnel to be designed, the feature value corresponding to the mileage of the tunnel to be designed is determined, and the feature value corresponding to the mileage of the tunnel to be designed is input into the target machine learning model to obtain a classification result sequence. The target light-dark boundary mileage of the tunnel to be designed is determined according to the classification result sequence. The feature values ​​include a first feature value, a second feature value, a third feature value, and a fourth feature value. The first feature value is the maximum elevation difference between the longitudinal section topographic line and the tunnel design axis on the longitudinal profile. The second feature value is the elevation difference between the cross-sectional topographic line and the top of the tunnel entrance on the cross-section. The third feature value is the cosine of the inclination angle corresponding to the straight line where the first and second intersection points are located on the cross-section. The first intersection point is the intersection of the first straight line and the topographic line. The first straight line is the straight line after rotating the left bottom vertical line of the tunnel entrance by a preset angle counterclockwise. The second intersection point is the intersection of the second straight line and the topographic line. The second straight line is the straight line after rotating the right bottom vertical line of the tunnel entrance by a preset angle clockwise. The fourth feature value is the ratio of the tunnel cross-sectional area below the topographic line on the cross-section to the total tunnel cross-sectional area. The step of determining the feature values ​​of the constructed tunnel entrance based on the dataset and according to a preset feature construction strategy includes: Based on the data of dense terrain points, determine the cross-sectional and longitudinal terrain lines at the entrances of constructed tunnels; The first feature value is determined based on the longitudinal profile topographic lines and the route data; The second feature value is determined based on the cross-sectional topographic lines, the route data, and the tunnel structure dimensions; The third characteristic value is determined based on the cross-sectional topographic lines and the tunnel structure dimensions. The fourth feature value is determined based on the cross-sectional topographic lines and the tunnel structural dimensions.

2. The method as described in claim 1, characterized in that, The steps of determining the feature values ​​of the constructed tunnel entrances based on the dataset according to a preset feature construction strategy, and generating positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage, include: A Delaunay triangulation object is created based on the first horizontal and vertical axis coordinates of the discrete point cloud in the terrain point cloud data. Generate the second horizontal and vertical coordinates of the dense point cloud, and determine whether the dense point cloud is inside the Delaunay triangulation object based on the second horizontal and vertical coordinates; When a dense point cloud is inside the Delaunay triangulation object, the vertical axis coordinates of the dense point cloud are determined by linear interpolation. When the dense point cloud is not inside the Delaunay triangulation object, the nearest discrete point cloud of the dense point cloud is determined according to the second horizontal and vertical axis coordinates, and the first vertical axis coordinate of the nearest discrete point cloud is used as the second vertical axis coordinate of the dense point cloud. Determine the terrain density point data based on the second horizontal and vertical axis coordinates and the second vertical axis coordinates; A new dataset of tunnel entrances that have been constructed is determined based on the terrain density point data, the route data, the tunnel size data, and the light-dark boundary mileage. Based on the new dataset, the feature values ​​of the tunnel entrances that have been constructed are determined according to a preset feature construction strategy, and positive and negative sample datasets are generated based on the feature values ​​and the light-dark boundary mileage.

3. The method as described in claim 1, characterized in that, The step of determining the first feature value based on the longitudinal profile topographic line and the route data includes: Based on the route data, the tunnel design axis, design start coordinates, and design end coordinates are determined, and multiple discrete points on the tunnel design axis are determined. Based on the design start coordinates and the design end coordinates, determine the three-dimensional coordinates of the discrete points on the axis; The first feature value is determined based on the three-dimensional coordinates and the longitudinal profile topographic lines.

4. The method as described in claim 1, characterized in that, The step of determining the second feature value based on the cross-sectional topographic line, the route data, and the tunnel structure dimensions includes: Based on the cross-sectional topographic lines, the route data, and the tunnel structural dimensions, determine the cross-sectional mileage, the design starting mileage, the design starting coordinates, the design ending coordinates, and the tunnel radius. Based on the cross-section mileage, design starting mileage, design starting coordinates, and design ending coordinates, the target coordinates are determined, wherein the target coordinates are the coordinates of the centerline of the cross-section tunnel; The terrain height value is determined based on the horizontal and vertical axis coordinates of the target coordinates, and the second feature value is determined based on the terrain height value, the vertical axis coordinate of the target coordinates, and the tunnel radius.

5. The method as described in claim 2, characterized in that, The step of determining the third characteristic value based on the cross-sectional topographic line and the tunnel structural dimensions includes: Determine the target unit normal vector of the longitudinal section where the tunnel axis is located, and determine the left bottom point and right bottom point of the tunnel entrance based on the target unit normal vector and the tunnel structure dimensions; Along the direction of the target unit normal vector, starting from the bottom left point of the tunnel entrance, move left or right by a preset step length to obtain the first current point. Determine the first height difference between the first straight line and the cross-sectional terrain line based on the horizontal and vertical coordinates of the first current point. Within a preset search range, continuously move and update the first current point based on the preset step length until the first height difference determined based on the first current point is less than a preset threshold. Then, determine the vertical coordinate of the first current point based on the horizontal and vertical coordinates of the first current point and the terrain dense point data, and use the horizontal and vertical coordinates of the first current point as the first intersection point. Along the direction of the target unit normal vector, with the bottom right point of the tunnel entrance as the second initial point, move left or right by a preset step length to obtain the second current point. Determine the second height difference between the second straight line and the cross-sectional terrain line based on the horizontal and vertical coordinates of the second current point. Within a preset search range, continuously move and update the second current point based on the preset step length until the second height difference determined based on the second current point is less than a preset threshold. Then, determine the vertical coordinate of the second current point based on the horizontal and vertical coordinates of the second current point and the terrain dense point data, and use the horizontal and vertical coordinates of the second current point as the second intersection point. The target vector is determined based on the first intersection point and the second intersection point, and the cosine of the angle between the target vector and the target unit normal vector is determined as the third eigenvalue.

6. The method as described in claim 1, characterized in that, The step of determining the fourth feature value based on the cross-sectional topographic line and the tunnel structural dimensions includes: Starting from the center of the tunnel entrance, the tunnel section is divided into units at preset intervals to the left and right along the radius of the horizontal direction of the tunnel entrance on the cross section. Calculate the height difference between the tunnel line and the terrain line in the cross section within each unit segment based on the tunnel structure dimensions. The area between the cross-sectional tunnel line and the cross-sectional topographic line is determined based on the height difference, and the area of ​​the overlapping region is determined based on the area. The fourth feature value is determined based on the area of ​​the overlapping region and the cross-sectional area of ​​the tunnel.

7. The method as described in claim 1, characterized in that, The positive and negative sample datasets include a positive sample dataset and a negative sample dataset; wherein, the step of generating the positive and negative sample datasets based on the feature values ​​and the light-dark boundary mileage includes: The feature values ​​located at the light-dark boundary mileage are identified as positive samples, and the positive sample dataset is constructed based on the positive samples; Feature values ​​far from the boundary between light and dark are identified as negative samples, and the negative sample dataset is constructed based on these negative samples.

8. The method as described in claim 7, characterized in that, The step of training a machine learning model based on the positive and negative sample dataset to obtain the target machine learning model includes: Set the range of values ​​for hyperparameters in the machine learning model; On the training set of the positive and negative sample datasets, the optimal hyperparameters are determined from the range of values ​​based on five-fold cross-validation; The machine learning model is retrained on the positive and negative sample datasets using the optimal hyperparameters to obtain the target machine learning model.

9. The method as described in claim 1, characterized in that, The classification result sequence includes multiple classification results, each including mileage, category, and probability; wherein, the step of determining the target light-dark boundary mileage of the tunnel to be designed based on the classification result sequence includes: The classification result with a preset value in the classification result sequence is determined as the first classification result; The probability ranking number of the first classification result is determined based on the probability of the classification result, and the first classification result whose probability ranking number is less than or equal to the preset ranking value is determined as the second classification result; The mileage furthest from the tunnel entrance of the tunnel to be designed in the second classification results is determined as the target light-dark boundary mileage.

Images