Subway station flood prevention simulation method and system based on three-dimensional model

Abstract

The invention relates to the technical field of twin models, in particular to a subway station flood prevention simulation method and system based on a three-dimensional model. According to the method, local important evaluation indexes are obtained through placename approximate analysis and concavity-convexity analysis of structures of point cloud data in a neighborhood space, and structural important evaluation indexes of single point cloud data are determined by combining and analyzing density distribution complexity in the neighborhood space; clustering the point cloud data and dividing the point cloud data into different regions; analyzing depth distribution of all the point cloud data in the regions; combining structural importance evaluation, determining an attention rate of each region, performing differential modeling and semantic processing to obtain a semantic three-dimensional model; and combining a hydrodynamic model to obtain a digital twinborn model. According to the method, differentiated three-dimensional modeling is carried out by evaluating the importance of each point cloud data in subway station flood prevention simulation, and flood prevention analysis is carried out more comprehensively and accurately by combining a digital twin model with a hydrodynamic model, so that the protection capability of subway flood prevention is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of twin model, in particular to a subway station flood prevention simulation method and system based on a three-dimensional model. BACKGROUND

[0002] As an underground closed space, subway will cause operational paralysis, equipment damage and even casualties and other disastrous consequences when flood or rainwater backflow occurs. Therefore, mechanical flood prevention monitoring and early warning are extremely important in order to ensure safety and maintain stable operation. Subway station flood prevention simulation can simulate the flooding process under different intensity of flood invasion by constructing a "digital twin" model of the subway station in a virtual space, thereby assisting in accurately identifying risk points, testing drainage capacity and optimizing emergency plans to provide intuitive basis for scientific decision-making and ensure the safety level of the subway system.

[0003] In general three-dimensional model generation, point cloud data obtained by three-dimensional laser scanning is usually processed uniformly to generate a model. However, uniform processing will treat data modeling without distinction, resulting in high-precision modeling of non-critical areas in the station, wasting a large amount of computing power and storage resources. Secondly, uniform allocation of precision will cause the elevation error of key flood prevention parts such as drainage ditch grating or entrance threshold to be covered by redundant details of non-critical areas, which may produce misleading safety signals and affect the accuracy of subsequent real simulation. SUMMARY

[0004] In order to solve the technical problem that uniform processing will treat data modeling without distinction, resulting in high-precision modeling of non-critical areas in the station, which will cause the key flood prevention parts to be covered by redundant details of non-critical areas and affect the accuracy of subsequent real simulation, the purpose of the present application is to provide a subway station flood prevention simulation method and system based on a three-dimensional model, and the technical solution adopted is as follows: The present application provides a subway station flood prevention simulation method based on a three-dimensional model, which comprises: obtaining point cloud data in the subway station to be tested; obtaining a local importance evaluation index of each point cloud data according to the ground approximate feature and the concave-convex feature of the structure distribution of each point cloud data in the neighborhood space; determining the structure importance evaluation index of each point cloud data based on the density distribution complexity of the point cloud data in the neighborhood space of each point cloud data and combining the local importance evaluation index; obtaining a region based on the neighborhood space distribution feature of the point cloud data; obtaining the attention rate of each region according to the depth distribution feature of the spatial position of each region and the structure importance evaluation index; constructing a semantic three-dimensional model based on the attention rate of different regions; The flood prevention digital twin model is obtained by combining the semantic three-dimensional model and the hydrodynamic model.

[0005] Further, the method for obtaining the local importance evaluation index comprises: For any point cloud data, all point cloud data in the neighborhood space of the point cloud data are taken as a neighborhood point set, and a covariance matrix of coordinates of the point cloud data in the neighborhood point set is obtained; after eigenvalue decomposition of the covariance matrix, a feature vector corresponding to the minimum eigenvalue is taken as a normal vector of the point cloud data; According to the deviation degree of the normal vector of the point cloud data from the vertical axis and the spatial bottom layer distribution, a near-ground importance value of the point cloud data is obtained; The curvature of the point cloud data is obtained based on the neighborhood point set of the point cloud data; and a value obtained by normalizing the absolute value of the curvature is taken as a concave-convex importance value of the point cloud data; The local importance evaluation index is obtained by combining the near-ground importance value and the concave-convex importance value of the point cloud data.

[0006] Further, the method for obtaining the near-ground importance value comprises: The mean value of the coordinate values of all point cloud data on the vertical axis is taken as a depth median value; and the difference between the depth median value and the coordinate value of the point cloud data on the vertical axis is normalized to obtain a distance near-ground index of the point cloud data; The direction near-ground index is obtained according to the size of the angle between the direction of the normal vector of the point cloud data and the direction of the vertical axis, and the angle and the direction near-ground index are negatively correlated; The near-ground importance value of the point cloud data is obtained by combining the direction near-ground index and the distance near-ground index of the point cloud data.

[0007] Further, the method for obtaining the structure importance evaluation index comprises: For any point cloud data, the number of other point cloud data with a shortest distance from the point cloud data to each preset direction less than a preset deviation value is counted as a point cloud distribution degree of each preset direction in the neighborhood space of the point cloud data; The density complexity index of the point cloud is obtained according to the non-zero chaotic degree of the point cloud distribution degrees of all preset directions; The structure importance evaluation index is obtained by combining the density complexity index and the local importance evaluation index.

[0008] Further, the method for obtaining a region by clustering based on the neighborhood space distribution characteristics of the point cloud data comprises: Any point cloud data is taken as a growth point, and region growing is performed according to a growth criterion to obtain a region; after the growing is completed, any point cloud data outside the region is taken as a new growth point to perform iterative region growing, and the growing is stopped when there is no new growth point for region growing; The growth criterion is: growth can be performed when the difference between the normal vector of the point cloud data to be grown and the growth point is less than the preset growth threshold.

[0009] Furthermore, the method for obtaining the attention rate includes: The coordinate values ​​of each point cloud data on the vertical axis are negatively correlated and normalized to obtain the depth of attention for each point cloud data. For any point cloud data within a region, the ratio of the depth of attention of that point cloud data to the sum of the depth of attention of all point cloud data within that region is used as the importance coefficient of that point cloud data. The importance coefficient is used as a weight to perform a weighted summation and normalization of the structural importance evaluation indicators of all point cloud data in the region, thereby obtaining the attention rate of the region.

[0010] Furthermore, the construction of a semantic 3D model based on attention rates in different regions includes: Areas with a attention rate greater than or equal to a preset high attention threshold are designated as high-importance areas; Regions with an attention rate lower than the preset high attention threshold but higher than the preset low attention threshold are designated as medium importance regions. Areas with a attention rate less than or equal to a preset low attention threshold are designated as low importance areas. Differentiated 3D models are constructed based on high-importance, medium-importance, and low-importance regions. High-importance regions are assigned high physical attributes, medium-importance regions are assigned basic physical attributes, and low-importance regions are not assigned physical attributes.

[0011] Furthermore, the process of combining the semantic 3D model with the hydrodynamic model to obtain a flood control digital twin model includes: A computational network was created by using a mesh generation tool to divide the semantic 3D model, mapping physical properties to hydrodynamic parameters, setting boundary conditions, and obtaining a flood control digital twin model through model coupling computation.

[0012] Furthermore, the neighborhood space is the size of a sphere with a preset radius, centered on each point cloud data point in three-dimensional space.

[0013] The present invention also provides a subway station flood control simulation system based on a three-dimensional model, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the subway station flood control simulation method based on a three-dimensional model as described above.

[0014] The present invention has the following beneficial effects: This invention utilizes toponymic approximation and concavity / convexity analysis of point cloud data in its neighborhood space to reflect the degree of influence of local point cloud data structures on flood control based on the potential water-bearing capacity and flatness of the local areas, thus providing a preliminary assessment of the importance of the point cloud data. Furthermore, considering the potential for seepage or drainage, the spatial density changes of point cloud data in key flood control areas are complex and dramatic. Therefore, the complexity of density distribution in the neighborhood space is analyzed, and combined with local importance assessment indicators, structural importance assessment indicators for individual point cloud data are determined to reflect the ultimate importance of single point cloud data for subway station flood control simulation. Since differentiated modeling requires operation on the entire area, the point cloud data is further clustered into different regions. The depth distribution of all point cloud data within each region is analyzed, and combined with structural importance assessment, the attention rate of each region is determined. Differential modeling and semantic processing are then performed to obtain a semantic 3D model, resulting in different construction accuracies and assigned different physical properties for different influencing parts. This facilitates the integration with hydrodynamic models, leading to a more efficient and accurate digital twin model with flood control analysis capabilities. This invention differentiates 3D modeling by assessing the importance of each point cloud data in the flood control simulation of subway stations, and combines it with a hydrodynamic model to generate a digital twin model with flood control analysis capabilities. It can dynamically deduce and visualize flood control scenarios, conduct more comprehensive and accurate flood control analysis, and improve the flood protection capabilities of subways. Attached Figure Description

[0015] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 A flowchart of a subway station flood control simulation method based on a three-dimensional model, provided as an embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for obtaining locally important evaluation indicators according to an embodiment of the present invention. Detailed Implementation

[0017] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a subway station flood control simulation method and system based on a three-dimensional model proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. Furthermore, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0018] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0019] The following description, in conjunction with the accompanying drawings, details the specific scheme of a subway station flood control simulation method and system based on a three-dimensional model provided by this invention.

[0020] Please see Figure 1 The diagram illustrates a flowchart of a subway station flood control simulation method based on a three-dimensional model, according to an embodiment of the present invention. The method includes the following steps: S1: Obtain point cloud data within the subway station to be tested.

[0021] In this embodiment of the invention, scanning points are strategically placed within the subway station and around its entrances and exits, and a 3D laser scanner is used to perform omnidirectional measurements of the subway station. The scanner projects laser beams onto the surfaces of surrounding objects through a high-speed rotating laser emitter, thereby acquiring several 3D coordinate points and obtaining dense point cloud data. It should be noted that the 3D coordinate system includes a horizontal axis and a vertical axis parallel to the Earth's surface, as well as a vertical axis perpendicular to the ground.

[0022] Since subway stations require multiple scanning points, each point cloud data point obtains coordinate data based on the scanner's temporary origin. To generate a complete subway station model from all point cloud data points, after acquisition, the point cloud data is matched using targets or feature points to unify the point cloud data from all scanning points into the same coordinate system. The final 3D coordinate origin can be a corner of a subway station entrance / exit. It should be noted that the point cloud data acquisition and unification process uses techniques well-known to those skilled in the art; the matching process will not be elaborated upon here. The coordinate system setting can be adjusted by the implementer according to the specific implementation scenario, such as directly using a geodetic coordinate system, etc., without restriction.

[0023] S2: Based on the approximate ground features and concavity / convexity features of the structural distribution of each point cloud data in the neighborhood space, obtain the local importance evaluation index for each point cloud data; based on the density distribution complexity of each point cloud data in the neighborhood space, combined with the local importance evaluation index, determine the structural importance evaluation index for each point cloud data.

[0024] When water enters a subway station, the water mainly accumulates and spreads on horizontal surfaces such as floors and platforms. The geometry of these surfaces, such as slope, depressions, or protrusions, will affect the depth, extent, and flow velocity of the accumulated water. Therefore, by analyzing the local geometric features of point cloud data, we can preliminarily assess the importance of point cloud data in influencing water flow carrying capacity.

[0025] In this embodiment of the invention, the proximity of each point cloud data is specifically analyzed through a preset neighborhood space. The neighborhood space is the size of a sphere with a preset radius length centered on each point cloud data in three-dimensional space. The preset radius length can be set to 5 centimeters. The specific value can be adjusted by the implementer and is not limited here.

[0026] The closer the local surface containing the point cloud data is to the ground plane, the more likely it is to be a primary carrier of water flow, and thus its importance increases. Furthermore, the unevenness or undulation of the local surface directly reflects water flow collection, storage, or obstruction and diversion, having a greater impact on hydrodynamics and thus increasing its importance. Therefore, a preliminary assessment of importance is conducted by combining approximate ground features and unevenness / undulation characteristics. Please refer to [link to relevant documentation]. Figure 2 The diagram illustrates a method for obtaining a locally important evaluation index according to an embodiment of the present invention. The method includes the following steps: S201: For any point cloud data, take all point cloud data in the neighborhood space as the neighborhood point set, and obtain the covariance matrix of the coordinates of the point cloud data in the neighborhood point set; after performing eigenvalue decomposition on the covariance matrix, take the eigenvector corresponding to the smallest eigenvalue as the normal vector of the point cloud data; based on the degree of deviation of the normal vector of the point cloud data from the vertical axis and the spatial bottom layer distribution, obtain the near-ground importance value of the point cloud data.

[0027] The set of points in the neighborhood space reflects the local surface conditions of the point cloud data. The closer the normal vector of the plane containing the point cloud data is to the vertical axis, the more likely the plane is to be parallel to the ground, and the higher the probability that it is a water-bearing plane. The covariance matrix is ​​used to characterize the relationships between the dimensions in a multidimensional dataset, and the eigenvectors of eigenvalue decomposition reflect the main direction of the data distribution. Therefore, the eigenvector with the smallest eigenvalue reflects the direction in which the coordinate changes of the point cloud data are least drastic, which is also the normal direction of the surface.

[0028] Considering the influence of the subway station ceiling, the point cloud data in the ceiling is unlikely to participate in water flow, and its importance is far less than that of the point cloud data on the ground. Therefore, in this embodiment of the invention, the average value of the coordinates of all point cloud data on the vertical axis is taken as the depth median. By analyzing the values ​​of the overall point cloud data on the vertical axis, the central position of the overall coordinate distribution is quantified. The lower the point cloud data is to the depth median, the higher its importance in water flow analysis.

[0029] The difference between the median depth and the coordinate value of the point cloud data on the vertical axis is then normalized and used as a near-ground indicator. The larger the difference, the more likely the point cloud data is to be located on the ground, and the higher its importance. It should be noted that normalization is a technique well-known to those skilled in the art. The choice of normalization can be linear normalization or standard normalization, such as the hyperbolic tangent function, etc. The specific normalization method is not limited here.

[0030] Furthermore, based on the angle between the normal direction and the vertical axis of the point cloud data, a near-ground orientation index is obtained. The angle is negatively correlated with the near-ground orientation index. The closer the normal direction is to the vertical axis, that is, the smaller the angle, the more parallel the local surface where the point cloud data is located is to the ground, and the higher its importance. In this embodiment of the invention, the angle between the normal direction and the vertical axis of the point cloud data is negatively correlated and normalized to obtain the near-ground orientation index.

[0031] It should be noted that negative correlation mapping is a technique well-known to those skilled in the art, such as using an inverse proportional function or a negative exponential function with the natural constant as the base, etc., and is not limited here. In a specific embodiment of the present invention, the expression for the direction approximation index is: In the formula, Represented as the first The directional near-ground indicator of point cloud data. Represented as the first The angle between the normal vector direction of a point cloud data and the vertical axis direction. It is represented as a normalization function.

[0032] Finally, by combining the directional near-ground index and the distance near-ground index of the point cloud data, the near-ground importance value of the point cloud data is obtained. In this embodiment of the invention, the product of the directional near-ground index and the distance near-ground index of the point cloud data is normalized to obtain the near-ground importance value of the point cloud data. The larger the directional near-ground index and the distance near-ground index, the more likely the point cloud data is to be a plane at the bottom that carries water flow, and the higher its importance.

[0033] S202: Analyze the curvature of the point cloud data in the neighborhood space to obtain the concavity and convexity importance of the point cloud data.

[0034] For the local geometric features of point cloud data, high positive curvature represents local depressions, reflecting that the local surface may be the bottom of a local depression, ditch, or drainage ditch, which has certain water flow convergence characteristics and should be given more attention in flood control simulation. At the same time, high negative curvature represents local protrusions, reflecting that the local surface may be the top of a protruding structure such as a threshold, step edge, or equipment foundation, which has a boundary interception effect on the inflow of water and also needs more attention in flood control simulation.

[0035] Therefore, in this embodiment of the invention, the curvature of the point cloud data is obtained based on the neighborhood point set of the point cloud data. It should be noted that curvature acquisition is a technique well-known to those skilled in the art, such as Gaussian curvature or principal component analysis, and will not be elaborated here. The value after normalizing the absolute value of the curvature is used as the concavity / convexity importance value of the point cloud data. The larger the curvature, the more significant the local concavity / convexity features, the greater the importance of participating in hydrodynamic influence, and the higher the attention required.

[0036] S203: Combine the near-ground importance value and the concavity / convexity importance value of the point cloud data to obtain the local importance assessment index.

[0037] By approximating the local surface with the ground plane and considering the local structural features, the importance is initially assessed. In this embodiment of the invention, the product of the near-ground importance value and the concavity / convexity importance value of the point cloud data is normalized to obtain a local importance assessment index. The importance is assessed through local geometric feature analysis. The higher the near-ground importance value and the concavity / convexity importance value, the higher the demand for attention, and the greater the importance of the point cloud data in the flood control simulation of the subway station.

[0038] Furthermore, when point cloud data exhibits uneven spatial density distribution, it can be used to measure the degree of complex geometric construction on the surface where the point cloud data resides. Compared to flat ground, the metal grilles of some drainage grates, the grooves in anti-slip tiles, ground cracks, and the edges of cable covers will cause multiple reflections during laser scanning, resulting in a dense distribution of point cloud data. This complex structure may also affect water infiltration or flow resistance, and its key details need to be preserved more completely during subsequent model construction.

[0039] Therefore, by combining the complexity of spatial density distribution with local importance evaluation indicators, the importance of point cloud data can be more comprehensively evaluated, resulting in structural importance evaluation indicators. In this embodiment of the invention, the method for obtaining structural importance evaluation indicators includes: First, for any given point cloud data, in its neighborhood space, the number of other point cloud data whose shortest distance to each preset direction is less than a preset deviation value is counted. This number is used as the point cloud distribution degree for each preset direction. In this embodiment of the invention, the preset directions are set to one direction every 10 degrees, providing 36 preset directions for analyzing density distribution. The preset deviation value is set to 2 centimeters. It should be noted that the implementer can adjust the numerical setting according to the implementation scenario, and there are no restrictions here. That is, if there are other point cloud data in the neighborhood space whose shortest distance to the preset direction is less than 2 centimeters, then this point cloud data is considered to be distributed in the preset direction. The number of all point cloud data distributed in the preset direction is counted as the point cloud distribution degree.

[0040] The more complex the distribution of point cloud values ​​in multiple directions, the more likely the point cloud data is located at complex structures, and the more detailed it needs to be preserved during subsequent model construction. Furthermore, based on the degree of disorder of the non-zero point cloud distribution values ​​in all preset directions, a density complexity index for the point cloud is obtained. In this embodiment of the invention, a disorder analysis is performed on the numerical point cloud distribution values, and the standard deviations of all non-zero point cloud distribution values ​​are normalized to obtain the density complexity index for the point cloud data. A larger standard deviation indicates that the spatial density is more likely to undergo complex and drastic changes, thus attracting greater attention.

[0041] Based on the importance of the geometric features of the point cloud data itself, and by combining the density complexity index and the local importance evaluation index, a structural importance evaluation index is obtained. In this embodiment of the invention, the product of the density complexity index and the local importance evaluation index is normalized to obtain the structural importance evaluation index. The larger the structural importance evaluation index, the more important the point cloud data is for subsequent flood control simulation analysis.

[0042] S3: Cluster the neighborhood spatial distribution features of point cloud data to obtain regions; obtain the attention rate of each region based on the depth distribution features of each region in spatial location and important structural evaluation indicators; construct a semantic 3D model based on the attention rates of different regions.

[0043] Considering that differentiated modeling primarily targets the overall structure rather than scattered data points, preliminary clustering is performed on the point cloud data to approximate the spatial geometric distribution characteristics into a single region. In this embodiment, a region growing algorithm is used for clustering. Any point cloud data point is used as a growth point, and region growth is performed according to a growth criterion to obtain the region. After growth is complete, any point cloud data point outside the region is used as a new growth point for iterative region growing until no new growth point is available. The growth criterion is: growth can proceed when the difference in normal vectors between the point cloud data to be grown and the growth point is less than a preset growth threshold.

[0044] In one specific embodiment of the present invention, the preset growth threshold can be set to 15°. Growth can be performed when the angle between the point cloud data to be grown and the normal vector between the growth point is less than the preset growth threshold. It should be noted that the region growing algorithm is a well-known technique to those skilled in the art, and will not be described further here.

[0045] Within each region, water flow tends to flow into locations with lower distribution. Therefore, when analyzing the simplification of subsequent modeling based on the overall regional importance, the depth distribution of spatial locations should be considered to make the assessment of important distribution more accurate.

[0046] Preferably, in this embodiment of the invention, the method for obtaining the attention rate includes: The coordinate values ​​of each point cloud data on the vertical axis are negatively correlated and normalized to obtain the depth of attention for each point cloud data. The smaller the coordinate value of the point cloud data on the vertical axis, the lower the distribution depth, and the more important it is for flood control analysis.

[0047] Furthermore, for any point cloud data within a given region, the ratio between the depth of attention of that point cloud data and the sum of the depth of attention values ​​of all point cloud data within that region is used as the importance coefficient of that point cloud data. The importance weight of each point cloud data within the region is quantified by measuring the overall depth of attention of the region.

[0048] The importance coefficient is used as a weight to perform a weighted summation and normalization of the structural importance evaluation index of all point cloud data in the region, thus obtaining the attention rate of the region. The higher the importance coefficient, the greater the attention importance of the corresponding point cloud data in the region. Therefore, using the importance coefficient as the weight, the product of the importance coefficient and the structural importance evaluation index of each point cloud data in the region is used as the weighted evaluation value of each point cloud data. The sum of all weighted evaluation values ​​in the region is normalized to obtain the attention rate. The attention rate reflects the importance of each region in the modeling.

[0049] A higher attention rate indicates that the region is more likely to be related to the basic behavior of water flow, such as advection, flow direction, and inflow degree. Simplifying this could lead to errors in the simulation results and generate misleading signals, requiring more precise preservation of structural details. Conversely, a lower attention rate indicates that the region has a lower impact on water flow behavior, allowing for greater simplification in modeling and improved modeling efficiency.

[0050] Therefore, a semantic 3D model is constructed based on the attention rate of different regions. In this embodiment of the invention, regions with an attention rate greater than or equal to a preset high attention threshold are designated as high-importance regions, regions with an attention rate less than the preset high attention threshold but greater than a preset low attention threshold are designated as medium-importance regions, and regions with an attention rate less than or equal to a preset low attention threshold are designated as low-importance regions. Different regions are modeled differently through different levels of attention rate. In one specific embodiment of the invention, the preset high attention threshold must be greater than the preset low attention threshold. The preset high attention threshold can be set to 0.8, and the preset low attention threshold can be set to 0.3. The specific values ​​can be adjusted by the implementer.

[0051] Differentiated 3D models are constructed based on high-importance, medium-importance, and low-importance regions. High-importance regions are key areas for flood control analysis; in this embodiment, the mesh precision can be selected within the minimum precision range, such as 10 to 15 centimeters, to restore details. Medium-importance regions are areas relevant to flood control analysis; in this embodiment, the mesh precision can be moderate, such as 30 to 50 centimeters, to preserve key shapes. For low-importance regions, which have less correlation with flood control analysis, such as ceiling areas, the mesh precision can be higher, such as 100 centimeters, to improve modeling efficiency. It should be noted that the specific numerical settings can be adjusted by the implementer according to the specific scenario of the embodiment, and are not limited here.

[0052] To ensure seamless integration with the hydrodynamic model, physical properties are assigned based on their importance, and semantic processing is used to achieve the final simulation process. In one specific embodiment of this invention, high-importance regions are assigned high-level physical properties, including complete parameters such as Manning coefficient, flow capacity or water-blocking height, reflection intensity, slope and aspect, etc., which can be obtained from design drawings, material samples, and field test reports. Medium-importance regions are assigned basic physical properties, which are some core parameters such as Manning coefficient and slope, characterizing macroscopic morphological properties. Low-importance regions may not be assigned physical properties to reduce computational costs.

[0053] S4: Combine semantic 3D model with hydrodynamic model to obtain flood control digital twin model.

[0054] After generating the semantic 3D model, it needs to be combined with the hydrodynamic model for data conversion and physical parameter mapping. In this embodiment of the invention, a mesh generation tool is used to divide the semantic 3D model into a computational network, and the physical attributes are mapped to hydrodynamic parameters to achieve mesh mapping of identifiable parameters. Boundary conditions are set, such as rainfall intensity, inlet and outlet flood intrusion flow, or water level process line. Based on the model coupling, the hydrodynamic solver is started to perform calculations. Based on the initial state and boundary conditions, the water depth and flow velocity of each grid cell within each time step are solved step by step. The dynamic evolution of water depth, flow velocity, and diffusion path is presented in real time through 3D visualization, thus completing the construction of the flood control digital twin model.

[0055] In summary, this invention uses place name approximation and concavity / convexity analysis of point cloud data in its neighborhood space to reflect the degree of influence of local point cloud data structure on flood control based on the potential water-bearing capacity and flatness of the local area, thus providing a preliminary assessment of the importance of the point cloud data. Furthermore, considering the potential for seepage or drainage, the spatial density changes of point cloud data in key flood control areas are complex and dramatic. Therefore, the complexity of density distribution in the neighborhood space is analyzed, and combined with local importance assessment indicators, structural importance assessment indicators for individual point cloud data are determined to reflect the ultimate importance of single point cloud data for subway station flood control simulation. Since differentiated modeling requires operation on the entire area, the point cloud data is further clustered into different regions. The depth distribution of all point cloud data within each region is analyzed, and combined with structural importance assessment, the attention rate of each region is determined. Differential modeling and semantic processing are then performed to obtain a semantic 3D model, resulting in different construction accuracies and assigned different physical properties for different influencing parts. This facilitates the integration with hydrodynamic models, leading to a more efficient and accurate digital twin model with flood control analysis capabilities. This invention differentiates 3D modeling by assessing the importance of each point cloud data in the flood control simulation of subway stations, and combines it with a hydrodynamic model to generate a digital twin model with flood control analysis capabilities. It can dynamically deduce and visualize flood control scenarios, conduct more comprehensive and accurate flood control analysis, and improve the flood protection capabilities of subways.

[0056] The present invention also provides a subway station flood control simulation system based on a three-dimensional model, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the subway station flood control simulation method based on a three-dimensional model as described above.

[0057] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0058] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A simulation method for flood control in subway stations based on a three-dimensional model, characterized in that, The method includes: Acquire point cloud data within the subway station to be tested; Based on the approximate ground features and concavity / convexity features of the structural distribution of each point cloud data in the neighborhood space, local importance evaluation indicators for each point cloud data are obtained; based on the complexity of the density distribution of each point cloud data in the neighborhood space, combined with the local importance evaluation indicators, structural importance evaluation indicators for each point cloud data are determined. Regions are obtained by clustering based on the neighborhood spatial distribution characteristics of point cloud data; attention rate of each region is obtained based on the depth distribution characteristics of each region in spatial location and structural importance evaluation indicators; semantic 3D models are constructed based on the attention rates of different regions. By combining a semantic 3D model with a hydrodynamic model, a digital twin model for flood control is obtained.

2. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The methods for obtaining the locally important evaluation indicators include: For any point cloud data, take all point cloud data in the neighborhood space as the neighborhood point set, and obtain the covariance matrix of the coordinates of the point cloud data in the neighborhood point set; after performing eigenvalue decomposition on the covariance matrix, take the eigenvector corresponding to the smallest eigenvalue as the normal vector of the point cloud data. Based on the degree of deviation between the normal vector and the vertical axis of the point cloud data and the spatial layer distribution, the near-ground importance value of the point cloud data is obtained. The curvature of the point cloud data is obtained based on the neighborhood point set of the point cloud data; the value after normalizing the absolute value of the curvature is used as the concavity and convexity importance value of the point cloud data. By combining the near-ground importance value and the concavity / convexity importance value of the point cloud data, a local importance assessment index is obtained.

3. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 2, characterized in that, The method for obtaining the near-Earth importance value includes: The mean of the coordinate values ​​of all point cloud data on the vertical axis is used as the depth median; the difference between the depth median and the coordinate value of the point cloud data on the vertical axis is normalized and used as the near-ground indicator of the point cloud data. Based on the angle between the normal vector direction and the vertical axis direction of the point cloud data, the near-ground orientation index is obtained, and the angle is negatively correlated with the near-ground orientation index. By combining the orientation near-ground index and distance near-ground index of the point cloud data, the near-ground importance value of the point cloud data is obtained.

4. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The methods for obtaining the key evaluation indicators of the structure include: For any point cloud data, in the neighborhood space of the point cloud data, count the number of other point cloud data whose shortest distance to each preset direction is less than a preset deviation value, and use this as the point cloud distribution degree for each preset direction; Based on the non-zero disorder of the point cloud distribution in all preset directions, the density complexity index of the point cloud is obtained. By combining density complexity indicators and local importance assessment indicators, structural importance assessment indicators are obtained.

5. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The regions obtained by clustering based on the neighborhood spatial distribution features of point cloud data include: Using any point cloud data as a growth point, perform region growth according to the growth criteria to obtain the region; after the growth is completed, use any point cloud data outside the region as a new growth point to iterate the region growth until there are no new growth points to perform region growth and then stop the growth. The growth criterion is: growth can be performed when the difference between the normal vector of the point cloud data to be grown and the growth point is less than the preset growth threshold.

6. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The methods for obtaining the attention rate include: The coordinate values ​​of each point cloud data on the vertical axis are negatively correlated and normalized to obtain the depth of attention for each point cloud data. For any point cloud data within a region, the ratio of the depth of attention of that point cloud data to the sum of the depth of attention of all point cloud data within that region is used as the importance coefficient of that point cloud data. The importance coefficient is used as a weight to perform a weighted summation and normalization of the structural importance evaluation indicators of all point cloud data in the region, thereby obtaining the attention rate of the region.

7. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The construction of a semantic 3D model based on attention rates in different regions includes: Areas with a attention rate greater than or equal to a preset high attention threshold are designated as high-importance areas; Regions with an attention rate lower than the preset high attention threshold but higher than the preset low attention threshold are designated as medium importance regions. Areas with a attention rate less than or equal to a preset low attention threshold are designated as low importance areas. Differentiated 3D models are constructed based on high-importance, medium-importance, and low-importance regions. High-importance regions are assigned high physical attributes, medium-importance regions are assigned basic physical attributes, and low-importance regions are not assigned physical attributes.

8. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The method of combining a semantic 3D model with a hydrodynamic model to obtain a flood control digital twin model includes: A computational network was created by using a mesh generation tool to divide the semantic 3D model, mapping physical properties to hydrodynamic parameters, setting boundary conditions, and obtaining a flood control digital twin model through model coupling computation.

9. The method for simulating flood control in subway stations based on a three-dimensional model according to claim 1, characterized in that, The neighborhood space is the size of a sphere with a preset radius, centered on each point cloud data point in three-dimensional space.

10. A subway station flood control simulation system based on a three-dimensional model, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the subway flood control simulation method based on a three-dimensional model as described in any one of claims 1 to 9.

Images