A method and system for urban area planning, construction, management and operation based on three-dimensional digital twins

Abstract

The present invention discloses a method and system for planning, construction, management and operation of urban areas based on three-dimensional digital twins, wherein the method comprises: collecting multi-source data of urban areas including geographic raster data, BIM model data and mobile phone signaling data, and realizing coordinate alignment of the multi-source data by LM algorithm; generating feature data composed of a coding table defining urban features from the multi-source data through feature engineering, and constructing an octree based on the feature data; defining the leaf nodes of the octree as voxels, applying multi-scale convolution to each voxel to obtain supervoxels with spatiotemporal labels; constructing a constraint network based on the supervoxels and a construction, management or operation and maintenance specification constraint library; determining the risk areas of the urban area by calculating the spatiotemporal constraint strength of the constraint edges, and outputting a corresponding risk area list. The present invention solves the problems of multi-source data integration, real-time dynamic monitoring and risk prediction in the planning, construction, management and operation and maintenance of urban areas.

Description

Technical Field

[0001] The present invention belongs to the field of digital twin technology, and in particular relates to a method and system for urban area planning, construction, management, operation and maintenance based on three-dimensional digital twins. Background Art

[0002] Currently, urban area planning, construction, management, and operation and maintenance primarily rely on a separate management model based on 2D GIS and BIM models, static 3D modeling techniques, and decision-making methods driven by human experience. Traditional 2D planning methods rely on planar blueprints for spatial layout design, resulting in a lack of 3D spatial information and difficulties in multidisciplinary collaboration. While digital twin technology, based on a single data source, can construct 3D models, it cannot integrate dynamic spatiotemporal data such as mobile phone signaling, resulting in a lack of data support for dynamic management, such as resource scheduling. Existing 3D modeling systems often employ a fixed-level spatial index structure, making them ill-suited to the dynamic changes in urban characteristics and the need for spatiotemporal correlation analysis. Regarding regulatory constraint management, existing methods typically rely on static manual verification against regulatory clauses, lacking the ability to monitor spatiotemporal conflicts among factors such as machinery, materials, and personnel during the construction process in real time. Risk identification, on the other hand, relies heavily on discrete sensor data and expert judgment, failing to achieve systematic risk prediction based on spatial topology and spatiotemporal evolution.

[0003] The Chinese patent publication number CN118279506A discloses a satellite image-based three-dimensional urban planning visualization system, including a system structure consisting of an application layer, a GIS service layer, a data layer, and a physical layer. The system's processing method for three-dimensional urban planning is as follows: Sp1: Satellite image setting and data acquisition, the satellite image setting and data acquisition is to connect the application layer in the system with the satellite image, to collect images within a specified urban area, and the satellite image selects the Beidou system, and the satellite image uses oblique photography technology to collect urban data; Sp2: Collection of urban geographic data, the collection of urban geographic data is the GIS service layer using geographic information system technology to collect data on urban ground information, geographical conditions, etc.; Sp3: Collection of urban color information, the collection of urban color information is the system summary The application layer connects to drones to collect low-altitude, multi-directional panoramic images of the city; Sp4: Data integration and processing, which involves aggregating the data collected in steps Sp1-Sp3 to the data layer in the system for resource integration and data analysis; Sp5: Urban three-dimensional modeling, which involves the data layer in the system using real-scene three-dimensional modeling technology to perform urban three-dimensional modeling based on the data information obtained; Sp6: Improvement of the urban three-dimensional model, which involves the application layer in the system reading the existing urban BIM model library information and performing data comparison of the corresponding building information to improve the urban three-dimensional model; Sp7: Urban three-dimensional planning, which involves visual planning of urban areas based on the urban three-dimensional model modeled in step Sp6. The system lacks real-time feedback and integration of operation and maintenance data during the operation and management of urban areas, and is unable to provide support for urban operation and maintenance and real-time changes. Summary of the Invention

[0004] The purpose of the present invention is to provide a method and system for urban area planning, construction, management and operation based on three-dimensional digital twins, so as to solve the technical problems of multi-source data integration, real-time dynamic monitoring and risk prediction in the planning, construction, management and operation of urban areas.

[0005] The technical solutions of the present invention are as follows:

[0006] In one aspect, the present invention provides a method for urban area planning, construction, management, operation and maintenance based on three-dimensional digital twins, comprising the following steps:

[0007] Collect multi-source data from urban areas, including geographic raster data, BIM model data, and mobile phone signaling data, and use the LM algorithm to achieve coordinate alignment of multi-source data;

[0008] The multi-source data is used to generate feature data consisting of a coding table defining city features through feature engineering, and an octree is constructed based on the feature data;

[0009] The leaf nodes of the octree are defined as voxels, and multi-scale convolution is applied to each voxel to obtain supervoxels with spatiotemporal labels;

[0010] Constructing a constraint network based on supervoxels and a construction, management, or operation and maintenance specification constraint library, wherein nodes in the constraint network are defined as mechanical nodes, material nodes, and personnel nodes, and constraint edges are defined as resource dependency edges and spatial conflict edges;

[0011] The risk areas of urban areas are determined by calculating the spatiotemporal constraint strength of the constraint edges, and the corresponding risk area list is output.

[0012] Preferably, the spatiotemporal alignment of multi-source data is achieved by the LM algorithm as follows:

[0013] Select the target coordinate system in geographic raster data, BIM model data, and mobile phone signaling data. Group the points with the same name and construct the error equation based on the points with the same name:

[0014]

[0015] Where, For the The error value of the same-name points, ; The target coordinate system Coordinates of points of the same name; For geographic raster data, BIM model data, mobile phone signaling data Coordinates of points of the same name; is the rotation matrix; is the translation vector;

[0016] Solve the optimal error value through SVD decomposition to obtain the initial rotation matrix and initial translation vector;

[0017] Define the coordinate transformation function based on the initial rotation matrix and the initial translation vector;

[0018] Construct a coordinate transformation optimization function based on the coordinate transformation function:

[0019]

[0020] Where, For coordinates Coordinate transformation optimization function; is the scale factor; is the coordinate point; is the initial rotation matrix; is the initial translation vector;

[0021] The LM algorithm is used to solve the coordinate transformation optimization function to obtain the optimal rotation matrix and optimal translation vector. The optimal rotation matrix and optimal translation vector are used to convert geographic raster data, BIM model data, and mobile phone signaling data into the target coordinate system to achieve coordinate alignment of multi-source data.

[0022] Preferably, the multi-source data is generated by feature engineering to generate feature data consisting of a coding table defining city features, and the octree is constructed based on the feature data as follows:

[0023] The feature data includes mechanical features, material features, and personnel features, and a three-dimensional space containing the entire urban area is initialized as a root node;

[0024] Calculate the difference of feature data within a node. When the difference of feature data within a node is greater than a set threshold, the node is divided into 8 sub-nodes. Each sub-node stores the statistical value of the city characteristics in the corresponding three-dimensional space. The feature data difference calculation is specifically as follows:

[0025]

[0026] Where, For child nodes The difference of feature data reflects the degree of dispersion of feature distribution; is the child node index of the octree, ; For child nodes Urban characteristics index; For child nodes Neidi Characteristic statistical values ​​of city-like characteristics; For child nodes The mean of all city characteristics within the city.

[0027] When the octree meets any of the following conditions: the difference within the node is less than the set threshold, the preset maximum level is reached, or the node volume is less than the preset value, the node splitting of the octree is terminated and the construction of the octree is completed.

[0028] Preferably, applying multi-scale convolution to each voxel is specifically as follows:

[0029]

[0030] Where, is the coordinate index of the voxel in three-dimensional space; is the category of urban characteristics; is a time slice; is the city feature enhancement value after multi-scale convolution; is the number of urban characteristics; is an index of urban characteristics; For the The feature weights of city-like features are obtained by dynamic calculation of mutual information; For the city-like characteristics; For the The corresponding convolution kernel of city-like features;

[0031] Based on multi-scale convolution, a supervoxel with spatiotemporal labels is obtained, where the supervoxel includes coordinates, time slices, city feature statistics, and city feature enhancement values.

[0032] Preferably, the constraint network is constructed based on the supervoxel and the construction, management or operation and maintenance specification constraint library as follows:

[0033] When the urban feature statistics and urban feature enhancement values ​​of the mechanical features, material features, and personnel features in the supervoxel exceed the corresponding set thresholds, and the supervoxel continuous space area exceeding the threshold is larger than the preset area, the supervoxel center point in the area is extracted as the corresponding mechanical node, material node, and personnel node;

[0034] Based on the construction, management or operation and maintenance specification constraint library, the resource dependency edges between the three types of nodes are defined. By calculating the distance between each node, if the node distance is less than the preset distance threshold, the nodes are connected and the constraint edge is defined as a spatial conflict edge.

[0035] Preferably, the spatiotemporal constraint strength of the constraint edge is specifically:

[0036]

[0037] Where, For the construction, management or operation and maintenance phase node 、 the spatiotemporal strength of the constraint edges between them; Number the construction, management, or operation and maintenance phase; 、 are node numbers respectively; For nodes 、 The Euclidean distance between The spatial influence radius is adjusted according to different types of constraint edges; is the time decay coefficient, which controls the speed at which the constraint decays over time; For the construction, management or operation and maintenance phase planned time; For the construction, management or operation and maintenance phase The base time;

[0038] According to the calculated spatiotemporal constraint strength of the constraint edge, the initial supervoxel constraint strength is injected into the supervoxel where the associated node is located, which is expressed as:

[0039]

[0040] Where, is the coordinate index of the supervoxel; is the initial supervoxel constraint strength of the supervoxel;

[0041] The supervoxel constraint strength is solved by a partial differential equation with a fixed time step according to the initial supervoxel constraint strength:

[0042]

[0043] Where, is the updated supervoxel constraint strength; is the time step; is the spatial diffusion coefficient; is the Laplace operator, calculate Diffusion in space; is the time conduction coefficient; is the adjacent supervoxel index; The coordinates are The supervoxel constraint strength of is the spacing between supervoxels;

[0044] The supervoxel risk level is determined according to the preset supervoxel constraint strength threshold, and the improved DBSCAN algorithm is used to merge adjacent high-risk supervoxels to form a continuous risk area, and the corresponding risk area list is output.

[0045] Preferably, the improved DBSCAN algorithm is used to merge adjacent high-risk supervoxels to form a continuous risk region:

[0046] The spatial domain radius and the minimum number of domain points of the cluster are set based on the weighted sum of the spatial distance and feature distance of the supervoxel;

[0047] Insert high-risk supervoxels into the octree according to their spatial positions, select unvisited high-risk supervoxels, and calculate the number of voxels of adjacent nodes in the neighborhood of the leaf node where the high-risk supervoxel is located. If the number of voxels in the traversed neighborhood is greater than or equal to the minimum number of neighboring points, mark the high-risk supervoxel as a core point and expand the cluster;

[0048] For each core point, all points in its neighborhood are obtained. If the neighborhood contains other core points, its neighborhood is also added to the current cluster and the cluster is continued to expand. If the neighborhood contains boundary points, the boundary points are added to the cluster.

[0049] If there are supervoxels that are not classified into any cluster during the expansion process, they will be marked as noise points;

[0050] Clustering is terminated after all high-risk supervoxels are visited and no new clusters can be expanded, and the final clustering result is obtained.

[0051] On the other hand, the present invention provides an urban area planning, construction, management and operation maintenance system based on three-dimensional digital twins, including a data acquisition and processing module, an octree construction module, a supervoxel generation module, a constraint network construction module and a risk area determination module;

[0052] The data acquisition and processing module is used to collect multi-source data of urban areas, including geographic raster data, BIM model data, and mobile phone signaling data, and realizes coordinate alignment of multi-source data through the LM algorithm;

[0053] An octree construction module, configured to generate feature data consisting of a coding table defining city features from the multi-source data through feature engineering, and to construct an octree based on the feature data;

[0054] The supervoxel generation module is used to define the leaf nodes of the octree as voxels and apply multi-scale convolution to each voxel to obtain supervoxels with spatiotemporal labels;

[0055] A constraint network construction module is used to construct a constraint network based on supervoxels and a construction, management, or operation and maintenance specification constraint library. Nodes in the constraint network are defined as mechanical nodes, material nodes, and personnel nodes, and constraint edges are defined as resource dependency edges and spatial conflict edges.

[0056] The risk area determination module is used to determine the risk areas of urban areas by calculating the spatiotemporal constraint strength of the constraint edges and output the corresponding risk area list.

[0057] On the other hand, the present invention also provides an electronic device, comprising: a memory, a processor, and a computer program stored on the memory and runnable on the processor, wherein when the processor executes the computer program, it implements the urban area planning, construction, management, and operation method based on three-dimensional digital twins as described in any embodiment of the present invention.

[0058] On the other hand, the present invention also provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it implements the urban area planning, construction, management and operation method based on three-dimensional digital twins as described in any embodiment of the present invention.

[0059] Compared with the prior art, the present invention has the following technical effects:

[0060] 1. This paper uses a hybrid coordinate transformation method that combines the LM algorithm with SVD decomposition to effectively solve the problem of spatial reference unification of multi-source heterogeneous data, including geographic raster data, BIM models, and mobile phone signaling. It breaks through the excessive reliance of traditional registration methods on manual point selection. By optimizing the scale factor and jointly solving the coordinate transformation function, it significantly improves the alignment accuracy of cross-scale and cross-dimensional spatial data, laying a reliable data foundation for subsequent spatiotemporal analysis.

[0061] 2. This invention innovatively achieves adaptive spatial partitioning of urban features by dynamically constructing an octree based on feature differences. By introducing a feature statistical difference criterion and an automatic segmentation mechanism, it ensures precise control of spatial partitioning while maintaining a complete representation of urban functional characteristics. Compared to traditional fixed-level partitioning, this dynamic spatial indexing method is more adaptable to the spatiotemporal dynamics of urban area characteristics.

[0062] 3. This paper designs a spatiotemporal fusion convolution kernel for three-dimensional voxels. By dynamically calculating feature weights based on mutual information, this method achieves cross-scale fusion of multi-dimensional urban features, such as architectural morphology and pedestrian flow dynamics. This method effectively enhances the spatiotemporal representation capabilities of supervoxels, enabling a single supervoxel to both carry spatial location attributes and express feature evolution patterns in the temporal dimension, providing high-dimensional feature support for subsequent constraint network construction.

[0063] 4. This invention innovatively establishes a digital mapping relationship between the specification constraint library and three-dimensional spatial elements. Through the constraint mechanism of resource-dependent edges and spatial conflict edges, it dynamically reflects the spatiotemporal interaction between machinery, materials, and personnel during the construction process, providing a constraint network framework for risk identification.

[0064] 5. This method constructs a supervoxel constraint strength propagation equation based on partial differential equations. By introducing spatial diffusion coefficients and temporal transmission coefficients, it accurately characterizes the propagation path and temporal evolution of risk in three-dimensional space. Compared with traditional discrete risk assessment methods, this method effectively reveals the spatiotemporal correlations of risk factors, significantly improving the accuracy and consistency of risk area identification in complex urban environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is an overall flow chart of the urban area planning, construction, management and operation maintenance method based on three-dimensional digital twins described in the present invention. DETAILED DESCRIPTION

[0066] In order to make the objectives, technical solutions and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in combination with specific embodiments of the present application and with reference to the accompanying drawings.

[0067] Example 1

[0068] This embodiment provides a method for planning, construction, management and operation of urban areas based on three-dimensional digital twins. Figure 1 As shown, the following steps are included:

[0069] Collect multi-source data from urban areas, including geographic raster data, BIM model data, and mobile phone signaling data, and use the LM algorithm to achieve coordinate alignment of these data sources. Specifically, geographic raster data can be collected through drone aerial surveys or satellite remote sensing, BIM model data can be generated using architectural design software, and mobile phone signaling data can be collected through operator base station recording.

[0070] As a preferred implementation of this embodiment, the spatiotemporal alignment of multi-source data is achieved by using the LM algorithm as follows:

[0071] Select the target coordinate system in geographic raster data, BIM model data, and mobile phone signaling data. Group points with the same name, such as building corners and road intersections, and construct error equations based on the same-name points:

[0072]

[0073] Where, For the The error value of the same-name points, ; The target coordinate system Coordinates of points of the same name; For geographic raster data, BIM model data, mobile phone signaling data Coordinates of points of the same name; is the rotation matrix (3*3); is the translation vector (3*1).

[0074] The optimal error value is solved through SVD decomposition to obtain the initial rotation matrix and initial translation vector.

[0075] Define the coordinate transformation function based on the initial rotation matrix and the initial translation vector.

[0076] Construct a coordinate transformation optimization function based on the coordinate transformation function:

[0077]

[0078] Where, For coordinates Coordinate transformation optimization function; is the scale factor, which is usually 1 by default; is the coordinate point; is the initial rotation matrix; is the initial translation vector.

[0079] The LM algorithm is used to solve the coordinate transformation optimization function to obtain the optimal rotation matrix and optimal translation vector. The optimal rotation matrix and optimal translation vector are used to convert geographic raster data, BIM model data, and mobile phone signaling data into the target coordinate system to achieve coordinate alignment of multi-source data.

[0080] Furthermore, the LM algorithm is used to solve the coordinate transformation optimization function, and the optimal rotation matrix and optimal translation vector are obtained as follows:

[0081] S1: Calculate the initial residual:

[0082]

[0083] Where, is the initial residual; The target coordinate system points of the same name; For geographic raster data, BIM model data, mobile phone signaling data Coordinates of points of the same name;

[0084] S2: Construct the Jacobian matrix:

[0085]

[0086] Where, is the Jacobian matrix; is the residual matrix; To optimize the parameters, ;

[0087] S3: Calculate the step size for each set of points with the same name and the target coordinate point using the residual and Jacobian matrix:

[0088]

[0089] Where, is the step length; is the damping factor; is the identity matrix;

[0090] S4: Update optimization parameters according to step size: ;

[0091] S5: Adjust the damping factor and repeat steps S3-S5 until the step size is less than the set threshold to obtain the optimal rotation matrix and optimal translation vector.

[0092] To quickly locate the voxel to which any coordinate point belongs, avoiding a full-space search, the multi-source data in this embodiment is used through feature engineering to generate feature data consisting of a code table defining city characteristics. An octree is then constructed based on this feature data. The octree is dynamically updated; when new data is added, only the nodes on the corresponding branch path need to be updated.

[0093] As a preferred implementation method of this embodiment, the multi-source data generates feature data composed of a coding table that defines urban characteristics through feature engineering. The following are some feature extraction examples: geographic raster data uses edge detection algorithms to extract road networks, green space contours, and water areas, and uses road building area analysis to extract the proportion of impermeable layers, etc.; BIM model data uses sensor data analysis to extract mechanical structure displacement, and uses geometric analysis to extract building height, pipeline complexity, structural load-bearing capacity, etc.; mobile phone signaling data uses cluster analysis to extract daytime population density, nighttime population distribution, flow hotspots, etc., and the extracted feature data are uniformly standardized.

[0094] Constructing an octree based on feature data is as follows:

[0095] The feature data includes mechanical features, material features, and personnel features, and a three-dimensional space including the entire urban area is initialized as a root node.

[0096] Calculate the difference of feature data within the node. When the difference of feature data within the node is greater than the set threshold (the threshold can be dynamically adjusted according to the size of the level, such as , for the Layer threshold, is the adjustment coefficient, is the root node level threshold), the node is divided into 8 child nodes, each child node stores the city feature statistics in the corresponding three-dimensional space, and the feature data difference is calculated as follows:

[0097]

[0098] Where, For child nodes The difference of feature data reflects the degree of dispersion of feature distribution; is the child node index of the octree, ; For child nodes Urban characteristics index; For child nodes Neidi Characteristic statistical values ​​of city-like features (such as mean, maximum, etc.). For example, if the feature is building height, then It can be the average height of all buildings in the subnode; For child nodes The mean of all city characteristics within the city.

[0099] When the octree meets any of the following conditions: intra-node diversity is less than a set threshold, the node size is less than a preset maximum, or the node volume is less than a preset value, the node splitting is terminated, completing the octree construction. Specifically, the preset maximum level or minimum node volume value is adaptively adjusted in practice based on the size of the urban area being used.

[0100] The leaf nodes of the octree are defined as voxels, and multi-scale convolution is applied to each voxel to obtain supervoxels with spatiotemporal labels.

[0101] As a preferred implementation of this embodiment, multi-scale convolution is applied to each voxel as follows:

[0102]

[0103] Where, is the coordinate index of the voxel in three-dimensional space; is the category of urban characteristics; is a time slice; is the city feature enhancement value after multi-scale convolution; is the number of urban characteristics; is an index of urban characteristics; For the The feature weights of city-like features are obtained by dynamic calculation of mutual information; For the city-like characteristics; For the The corresponding convolution kernel of city-like features can preset convolution kernel parameters based on different features and urban planning experience. For example, in the application scenario of pipeline node analysis, the convolution kernel size can be set to 3*3*3 to extract local features; in the application scenario of traffic flow analysis, the convolution kernel size can be set to 5*5*5 to extract regional features.

[0104] Further, The calculation is expressed as:

[0105]

[0106] Where, Characterized by With the target variable Mutual information (classifying various features into different target variables, such as different pipeline complexity intervals for different construction risk levels).

[0107] Based on multi-scale convolution, a supervoxel with spatiotemporal labels is obtained, where the supervoxel includes coordinates, time slices, city feature statistics, and city feature enhancement values.

[0108] A constraint network is constructed based on supervoxels and a construction, management or operation and maintenance specification constraint library, wherein nodes in the constraint network are defined as mechanical nodes, material nodes and personnel nodes, and constraint edges are defined as resource dependency edges and spatial conflict edges.

[0109] As a preferred implementation of this embodiment, the constraint network is constructed based on supervoxels and a construction, management or operation and maintenance specification constraint library as follows:

[0110] When the urban feature statistics and enhancement values ​​for mechanical, material, and personnel features within a supervoxel exceed corresponding thresholds (e.g., the daytime population density representing the personnel feature within the supervoxel exceeds a set threshold), and a continuous region of supervoxels exceeding the threshold exists that is larger than a preset area, the supervoxel center within that region is extracted as the corresponding mechanical, material, and personnel nodes. Specifically, in this embodiment, node selection is based on the core factors affecting construction safety and progress, material nodes determining resource scheduling and space utilization efficiency, and personnel nodes involving area safety management.

[0111] Based on the construction, management, or operation and maintenance specification constraint library (such as the implementation regulations and specification manuals in the construction, management, or operation and maintenance process), the resource dependency edges between the three types of nodes are defined. By calculating the distance between each node, if the node distance is less than the preset distance threshold, the nodes are connected, and the constraint edge is defined as a spatial conflict edge.

[0112] The risk areas of urban areas are determined by calculating the spatiotemporal constraint strength of the constraint edges, and the corresponding risk area list is output.

[0113] As a preferred implementation of this embodiment, the spatiotemporal constraint strength of the constraint edge is specifically:

[0114]

[0115] Where, For the construction, management or operation and maintenance phase node 、 The spatiotemporal strength of the constraint edges between them reflects new progress and environmental changes; Number the construction, management, or operation and maintenance phase; 、 are node numbers respectively; For nodes 、 The Euclidean distance between The spatial influence radius is adjusted according to different types of constraint edges; is the time decay coefficient, which controls the speed at which the constraint decays over time; For the construction, management or operation and maintenance phase planned time; For the construction, management or operation and maintenance phase The base time.

[0116] According to the calculated spatiotemporal constraint strength of the constraint edge, the initial supervoxel constraint strength is injected into the supervoxel where the associated node is located, which is expressed as:

[0117]

[0118] Where, is the coordinate index of the supervoxel; is the initial supervoxel constraint strength of the supervoxel;

[0119] The supervoxel constraint strength is solved by a partial differential equation with a fixed time step according to the initial supervoxel constraint strength:

[0120]

[0121] Where, is the updated supervoxel constraint strength; is the time step; is the spatial diffusion coefficient; is the Laplace operator, calculate Diffusion in space, ; is the time conduction coefficient; is the adjacent supervoxel index; The coordinates are The supervoxel constraint strength of is the distance between supervoxels.

[0122] Within a single stage, the spatiotemporal strength values ​​of the constraint edges remain unchanged, and only the supervoxel constraint strength is updated through partial differential equations; when entering a new stage, the spatiotemporal strength values ​​of the constraint edges are recalculated and the supervoxel constraint strength is reinitialized.

[0123] The supervoxel risk level is determined according to the preset supervoxel constraint strength threshold, and the improved DBSCAN algorithm is used to merge adjacent high-risk supervoxels to form a continuous risk area, and the corresponding risk area list is output.

[0124] As a preferred implementation of this embodiment, the improved DBSCAN algorithm is used to merge adjacent high-risk supervoxels to form a continuous risk region. Specifically:

[0125] The spatial domain radius and the minimum number of domain points of the cluster are set based on the weighted sum of the spatial distance and feature distance of the supervoxel.

[0126] High-risk supervoxels are inserted into the octree according to their spatial positions to accelerate the neighborhood search without the need for a global search. Unvisited high-risk supervoxels are selected, and the number of voxels of adjacent nodes in the neighborhood of the leaf node where the high-risk supervoxel is located is calculated. If the number of voxels in the traversed neighborhood is greater than or equal to the minimum number of neighboring points, the high-risk supervoxel is marked as a core point and the cluster is expanded.

[0127] For each core point, all points in its neighborhood are obtained. If the neighborhood contains other core points, its neighborhood is also added to the current cluster and the cluster is continued to expand; if the neighborhood contains boundary points, the boundary points are added to the cluster.

[0128] If there are supervoxels that are not classified into any cluster during the expansion process, they are marked as noise points.

[0129] Clustering is terminated after all high-risk supervoxels have been visited and no new clusters can be expanded, resulting in the final clustering result. Furthermore, when a new high-risk supervoxel is added, only its neighborhood relationship with existing clusters needs to be checked, without recalculating the entire data.

[0130] Specifically, the risk area list includes clustered risk area results and corresponding solutions. The solutions are further generated by: pre-defining a measure library, selecting the optimal measures for matching based on cost and resource constraints, and the matching logic includes matching the scope of the measures with the geographical location of the high-risk areas and the dominant characteristics of the measures for the areas.

[0131] Example 2

[0132] Accordingly, this embodiment provides a system for urban area planning, construction, management, and operation based on three-dimensional digital twins. The system is used to implement the method for urban area planning, construction, management, and operation based on three-dimensional digital twins as described in Example 1 of the present invention, including a data acquisition and processing module, an octree construction module, a supervoxel generation module, a constraint network construction module, and a risk area determination module.

[0133] The data acquisition and processing module is used to collect multi-source data of urban areas, including geographic raster data, BIM model data, and mobile phone signaling data, and realizes coordinate alignment of multi-source data through the LM algorithm;

[0134] An octree construction module, configured to generate feature data consisting of a coding table defining city features from the multi-source data through feature engineering, and to construct an octree based on the feature data;

[0135] The supervoxel generation module is used to define the leaf nodes of the octree as voxels and apply multi-scale convolution to each voxel to obtain supervoxels with spatiotemporal labels;

[0136] A constraint network construction module is used to construct a constraint network based on supervoxels and a construction, management, or operation and maintenance specification constraint library. Nodes in the constraint network are defined as mechanical nodes, material nodes, and personnel nodes, and constraint edges are defined as resource dependency edges and spatial conflict edges.

[0137] The risk area determination module is used to determine the risk areas of urban areas by calculating the spatiotemporal constraint strength of the constraint edges and output the corresponding risk area list.

[0138] Example 3

[0139] This embodiment provides an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and runnable on the processor. When the processor executes the computer program, it implements the urban area planning, construction, management, and operation method based on three-dimensional digital twins as described in Example 1 of the present invention.

[0140] Example 4

[0141] This embodiment provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method for planning, construction, management, and operation of urban areas based on three-dimensional digital twins as described in the first embodiment of the present invention is implemented.

[0142] In the embodiments of the present application, "at least one" refers to one or more, and "more" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent the existence of A alone, the existence of A and B at the same time, and the existence of B alone. Among them, A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c can represent: a, b, c, a and b, a and c, b and c or a and b and c, where a, b, c can be single or multiple.

[0143] Those skilled in the art will appreciate that the various units and algorithm steps described in the embodiments disclosed herein can be implemented using a combination of electronic hardware, computer software, and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0144] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0145] In the several embodiments provided in this application, if any function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of this application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory; hereinafter referred to as: ROM), random access memory (Random Access Memory; hereinafter referred to as: RAM), magnetic disk or optical disk, and other media that can store program code.

[0146] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention's description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for urban area planning, construction, management and operation based on three-dimensional digital twins, characterized by: The following steps are involved: Collect multi-source data from urban areas, including geographic raster data, BIM model data, and mobile phone signaling data, and use the LM algorithm to achieve coordinate alignment of multi-source data; The multi-source data is used to generate feature data consisting of a coding table defining city features through feature engineering, and an octree is constructed based on the feature data; The leaf nodes of the octree are defined as voxels, and multi-scale convolution is applied to each voxel to obtain supervoxels with spatiotemporal labels; A constraint network is constructed based on supervoxels and a construction, management, or operation and maintenance specification constraint library. Nodes in the constraint network are defined as mechanical nodes, material nodes, and personnel nodes, and constraint edges are defined as resource dependency edges and spatial conflict edges. Specifically, resource dependency edges are defined between the three types of nodes based on the construction, management, or operation and maintenance specification constraint library. The distance between each node is calculated. If the distance between nodes is less than a preset distance threshold, the nodes are connected, and the constraint edge is defined as a spatial conflict edge. The risk areas of the urban area are determined by calculating the spatiotemporal constraint strength of the constraint edge and outputting a corresponding risk area list; the spatiotemporal constraint strength of the constraint edge is specifically: Where, For the construction, management or operation and maintenance phase node 、 the spatiotemporal strength of the constraint edges between them; Number the construction, management, or operation and maintenance phase; 、 are node numbers respectively; For nodes 、 The Euclidean distance between The spatial influence radius is adjusted according to different types of constraint edges; is the time decay coefficient, which controls the speed at which the constraint decays over time; For the construction, management or operation and maintenance phase planned time; For the construction, management or operation and maintenance phase The base time.

2. The urban area planning, construction, management and operation method based on three-dimensional digital twin according to claim 1 is characterized in that: The coordinate alignment of multi-source data is achieved through the LM algorithm as follows: Select the target coordinate system in geographic raster data, BIM model data, and mobile phone signaling data. Group the points with the same name and construct the error equation based on the points with the same name: Where, For the The error value of the same-name points, ; The target coordinate system Coordinates of points of the same name; For geographic raster data, BIM model data, mobile phone signaling data Coordinates of points of the same name; is the rotation matrix; is the translation vector; Solve the optimal error value through SVD decomposition to obtain the initial rotation matrix and initial translation vector; Define the coordinate transformation function based on the initial rotation matrix and the initial translation vector; Construct a coordinate transformation optimization function based on the coordinate transformation function: Where, For coordinates Coordinate transformation optimization function; is the scale factor; is the coordinate point; is the initial rotation matrix; is the initial translation vector; The LM algorithm is used to solve the coordinate transformation optimization function to obtain the optimal rotation matrix and optimal translation vector. The optimal rotation matrix and optimal translation vector are used to convert geographic raster data, BIM model data, and mobile phone signaling data into the target coordinate system to achieve coordinate alignment of multi-source data.

3. The urban area planning, construction, management and operation method based on three-dimensional digital twin according to claim 1 is characterized in that: The multi-source data is used to generate feature data consisting of a coding table defining city features through feature engineering. The octree is constructed based on the feature data as follows: The feature data includes mechanical features, material features, and personnel features, and a three-dimensional space containing the entire urban area is initialized as a root node; Calculate the difference of feature data within a node. When the difference of feature data within a node is greater than a set threshold, the node is divided into 8 sub-nodes. Each sub-node stores the statistical value of the city characteristics in the corresponding three-dimensional space. The feature data difference calculation is specifically as follows: Where, For child nodes The difference of feature data reflects the degree of dispersion of feature distribution; is the child node index of the octree, ; For child nodes Urban characteristics index; For child nodes Neidi Characteristic statistical values ​​of city-like characteristics; For child nodes The mean of all city characteristics within the When the octree meets any of the following conditions: the difference within the node is less than the set threshold, the preset maximum level is reached, or the node volume is less than the preset value, the node splitting of the octree is terminated and the construction of the octree is completed.

4. The urban area planning, construction, management and operation method based on three-dimensional digital twin according to claim 3 is characterized in that: Applying multi-scale convolution to each voxel is as follows: Where, is the coordinate index of the voxel in three-dimensional space; is the category of urban characteristics; is a time slice; is the city feature enhancement value after multi-scale convolution; is the number of urban characteristics; is an index of urban characteristics; For the The feature weights of city-like features are obtained by dynamic calculation of mutual information; For the city-like characteristics; For the The corresponding convolution kernel of city-like features; Based on multi-scale convolution, a supervoxel with spatiotemporal labels is obtained, where the supervoxel includes coordinates, time slices, city feature statistics, and city feature enhancement values.

5. The urban area planning, construction, management and operation method based on three-dimensional digital twin according to claim 4 is characterized in that: The nodes in the constraint network are defined as mechanical nodes, material nodes, and personnel nodes as follows: When the urban feature statistics and urban feature enhancement values ​​of the mechanical features, material features and personnel features in the supervoxel exceed the corresponding set thresholds, and there is a supervoxel continuous space area exceeding the threshold that is larger than the preset area, the supervoxel center point in the area is extracted as the corresponding mechanical node, material node and personnel node.

6. The urban area planning, construction, management and operation method based on three-dimensional digital twin according to claim 1 is characterized in that: Determine the risk areas of the urban area and output the corresponding risk area list as follows: According to the calculated spatiotemporal constraint strength of the constraint edge, the initial supervoxel constraint strength is injected into the supervoxel where the associated node is located, which is expressed as: Where, is the coordinate index of the supervoxel; is the initial supervoxel constraint strength of the supervoxel; The supervoxel constraint strength is solved by a partial differential equation with a fixed time step according to the initial supervoxel constraint strength: Where, is the updated supervoxel constraint strength; is the time step; is the spatial diffusion coefficient; is the Laplace operator, calculate Diffusion in space; is the time conduction coefficient; is the adjacent supervoxel index; The coordinates are The supervoxel constraint strength of is the spacing between supervoxels; The supervoxel risk level is determined according to the preset supervoxel constraint strength threshold, and the improved DBSCAN algorithm is used to merge adjacent high-risk supervoxels to form a continuous risk area, and the corresponding risk area list is output.

7. The urban area planning, construction, management and operation method based on three-dimensional digital twin according to claim 6 is characterized in that: The improved DBSCAN algorithm is used to merge adjacent high-risk supervoxels to form a continuous risk region: The spatial domain radius and the minimum number of domain points of the cluster are set based on the weighted sum of the spatial distance and feature distance of the supervoxel; Insert high-risk supervoxels into the octree according to their spatial positions, select unvisited high-risk supervoxels, and calculate the number of voxels of adjacent nodes in the neighborhood of the leaf node where the high-risk supervoxel is located. If the number of voxels in the traversed neighborhood is greater than or equal to the minimum number of neighboring points, mark the high-risk supervoxel as a core point and expand the cluster; For each core point, all points in its neighborhood are obtained. If the neighborhood contains other core points, its neighborhood is also added to the current cluster and the cluster is continued to expand. If the neighborhood contains boundary points, the boundary points are added to the cluster. If there are supervoxels that are not classified into any cluster during the expansion process, they will be marked as noise points; Clustering is terminated after all high-risk supervoxels are visited and no new clusters can be expanded, and the final clustering result is obtained.

8. An urban area planning, construction, management and operation system based on three-dimensional digital twins, characterized by: The system is used to implement the urban area planning, construction, management and operation maintenance method based on three-dimensional digital twins according to any one of claims 1 to 7, including a data acquisition and processing module, an octree construction module, a supervoxel generation module, a constraint network construction module and a risk area determination module; The data acquisition and processing module is used to collect multi-source data of urban areas, including geographic raster data, BIM model data, and mobile phone signaling data, and realizes coordinate alignment of multi-source data through the LM algorithm; An octree construction module, configured to generate feature data consisting of a coding table defining city features from the multi-source data through feature engineering, and to construct an octree based on the feature data; The supervoxel generation module is used to define the leaf nodes of the octree as voxels and apply multi-scale convolution to each voxel to obtain supervoxels with spatiotemporal labels; A constraint network construction module is used to construct a constraint network based on supervoxels and a construction, management, or operation and maintenance specification constraint library. Nodes in the constraint network are defined as mechanical nodes, material nodes, and personnel nodes, and constraint edges are defined as resource dependency edges and spatial conflict edges. The risk area determination module is used to determine the risk areas of urban areas by calculating the spatiotemporal constraint strength of the constraint edges and output the corresponding risk area list.

9. An electronic device, 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, the method for planning, construction, management, and operation of urban areas based on three-dimensional digital twins as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for planning, construction, management and operation of urban areas based on three-dimensional digital twins as described in any one of claims 1 to 7 is implemented.

Images