Comprehensive management and control method and system based on digital twin and multi-source heterogeneous data processing

Abstract

The present invention provides a comprehensive management and control method and system based on digital twins and multi-source heterogeneous data processing, the following steps: scan the site, obtain point cloud data of the surrounding environment, reconstruct the reverse modeling of the geometric shape and structure of the building based on the point cloud data, and calibrate the model obtained by reverse modeling with the site; optimize the model; use multi-source heterogeneous data fusion technology to process and classify the heterogeneous data of various professional rail transit, and realize standardized data processing. The comprehensive management and control method and system based on digital twins and multi-source heterogeneous data processing provided by the present invention realizes the accurate mapping of the digital twin space to the real station environment, can associate the three-dimensional model with the multi-dimensional data attributes, assist in realizing the core functions such as virtual inspection, risk control and scheduling decision-making, and improve the intelligent operation level of the station.

Description

Technical Field

[0001] The present invention belongs to the field of intelligent transportation technology, and specifically relates to a comprehensive management and control method and system based on digital twins and multi-source heterogeneous data processing. Background Art

[0002] As urban rail transit systems continue to expand and operating mileage continues to increase, traditional station operations and maintenance methods are facing challenges such as inefficiency, data silos, and difficulties in cross-system collaboration. Existing operations and maintenance management systems typically rely on individual management, manual inspections, and passive alerting, making it difficult to monitor and predict station operating status in real time. This not only increases operations and maintenance costs, but also hinders the accuracy and timeliness of dispatching and management decisions.

[0003] At present, most of the integrated management and control systems based on digital twins and multi-source heterogeneous data processing introduce single-professional or single-function digital and information systems for management. Different modules (such as equipment management, fault alarm, etc.) are often relatively independent; data cannot be shared, and it is difficult for systems to work together, and it is difficult to achieve cross-system data integration and utilization; lack of real-time data support, unable to achieve accurate positioning and fault prediction of equipment, unable to effectively improve the problems of operation and maintenance system fragmentation, data barriers, etc., resulting in a lack of necessary support for station management and decision support, and unable to adapt to the requirements of smart station operations.

[0004] The existing integrated management and control system based on digital twins and multi-source heterogeneous data processing has limited communication mechanisms, interface protocols, etc. Different business function modules are relatively independent, and cannot be coordinated and called, resulting in system fragmentation. At the same time, data sources from different systems lead to inconsistent formats, making it difficult to achieve effective integration and use of cross-system data. There is a lack of real-time data support for overall station situation awareness, and a lack of technical support for scientific and effective intelligent decision-making, resulting in the inability to adapt to more diverse station service needs. Summary of the Invention

[0005] The purpose of this invention is to provide a comprehensive management and control method and system based on digital twins and multi-source heterogeneous data processing, aiming to solve the technical problem that the existing comprehensive management and control system based on digital twins and multi-source heterogeneous data processing cannot adapt to the intelligent operation of stations.

[0006] To achieve the above objectives, the technical solution adopted by the present invention is to provide a comprehensive management and control method based on digital twins and multi-source heterogeneous data processing, including the following steps:

[0007] Scan the site to obtain point cloud data of the surrounding environment, reconstruct the building's geometry and reverse modeling based on the point cloud data, and calibrate the reverse modeling model with the site;

[0008] Optimize the model;

[0009] Multi-source heterogeneous data fusion technology is used to process the collection and classification of heterogeneous data from various rail transit disciplines to achieve standardized data processing.

[0010] Preferably, scanning the site to obtain point cloud data of the surrounding environment, reconstructing the geometric shape and structure of the building based on the point cloud data, and calibrating the model obtained by the reverse modeling with the site, includes:

[0011] Obtain a large amount of point cloud data through the 3D laser scanning process, perform denoising on the point cloud data, and perform point cloud registration operations;

[0012] The point cloud data is further converted into a three-dimensional model through surface fitting; the fitting formula is:

[0013]

[0014] in, represents the function of fitting the surface, are the surface parameters to be solved, represents a point in the point cloud, represents the fitting error.

[0015] Compare the model with the actual environment data to verify consistency; the comparison formula is:

[0016]

[0017] in, represents the consistency measure between the digital twin and the actual environment, represents the calibration matrix of the adjustment parameters, represents the model points generated by reverse modeling, Represents actual data points from a production environment.

[0018] Preferably, the process of acquiring a large amount of point cloud data through a three-dimensional laser scanning process, performing denoising on the point cloud data, and performing a point cloud registration operation includes:

[0019] The registration formula used is:

[0020]

[0021] Among them, T represents the rigid body transformation matrix including the rotation matrix and displacement vector, which is used to align the point cloud data. and represents corresponding points from two different point clouds, represents the Euclidean distance metric.

[0022] Preferably, the optimization processing of the model includes: using models of different degrees of refinement to represent the same model; wherein the level of detail is adjusted according to the distance from the model to the observer using LOD technology:

[0023]

[0024] in, Indicates the LOD level, Represents the distance from the model to the observer, represents the resolution factor, Indicates the proportion of the model on the screen;

[0025] Calculate the rendering complexity of the model:

[0026]

[0027] in, Indicates rendering complexity, Indicates the number of triangles in the model, Indicates the amount of rendering calculation for each triangle;

[0028] Calculate the scene rendering time after LOD technology:

[0029]

[0030] in, Indicates the total rendering time, Indicates the number of models in the scene, Indicates the The basic rendering time of each model, Indicates the The LOD level of the model.

[0031] Preferably, the optimization processing of the model includes: performing low-poly baking on the fine model to generate models of different precisions, wherein when in use, the selection of the fineness of the model depends on the distance between the viewpoint and the model.

[0032] Preferably, the optimizing the model includes:

[0033] Standardize heterogeneous data; the data standardization formula is:

[0034]

[0035] in, Represents the original data value, represents the mean of the data, Represents the standard deviation of the data. By normalizing the data to a uniform scale, it ensures that the data is compared and processed on the same dimension;

[0036] Through fusion algorithms, data from different systems are mapped, associated, and merged to ultimately form unified structured data;

[0037] Carry out process-based collection, classification and distribution of heterogeneous data from various rail transit disciplines, unify the mapping of three different coding systems: equipment coding, asset coding and component coding, and generate a unique code.

[0038] Preferably, the data of different systems are mapped, associated and merged through a fusion algorithm to finally form unified structured data. The fusion algorithm includes one or more of a key-value-based association matching algorithm and a semantic-based heterogeneous data matching algorithm.

[0039] Preferably, the process of collecting, classifying and distributing heterogeneous data for various rail transit disciplines is carried out, and the three different coding systems of equipment coding, asset coding and component coding are uniformly mapped to generate a unique code, including:

[0040] Clean and format the code to extract the valid part from the code string;

[0041] Standardize the three different codes: equipment code, asset code and component code;

[0042] Hash the encoding to generate a unique identifier of fixed length.

[0043] Preferably, the cleaning and formatting of the code to extract the valid part from the code string includes: using regular expression parsing to extract the matching string:

[0044]

[0045] in, is the original encoding, is a function that matches regular expressions. It is the normalized encoding.

[0046] The present invention also provides an integrated management and control system based on digital twins and multi-source heterogeneous data processing, characterized in that it is used to execute the integrated management and control method based on digital twins and multi-source heterogeneous data processing as described in any one of the above items, including:

[0047] The model conversion module is used to scan the site, obtain point cloud data of the surrounding environment, reconstruct the building's geometry and reverse modeling of the structure based on the point cloud data, and calibrate the model obtained by reverse modeling with the site;

[0048] Optimization module, used to optimize the model;

[0049] The fusion module is used to adopt multi-source heterogeneous data fusion technology to process and classify heterogeneous data from various rail transit disciplines, thereby achieving standardized data processing.

[0050] The beneficial effects of the comprehensive management and control method, system and assembly method based on digital twin and multi-source heterogeneous data processing provided by the present invention are as follows: compared with the prior art, the comprehensive management and control method and system based on digital twin and multi-source heterogeneous data processing of the present invention realize efficient data processing and spatial topological structure construction, realize lightweight and real-time rendering of large models, improve the smooth loading and display of large scenes and large-volume models, and ensure high-performance work in big data scenarios. The comprehensive management and control method based on digital twin and multi-source heterogeneous data processing standardizes data from different systems, realizes cross-system data sharing and calling, facilitates system data intercommunication, and improves data utilization efficiency. The comprehensive management and control method based on digital twin and multi-source heterogeneous data processing realizes the accurate mapping of the digital twin space to the real station environment, can associate the three-dimensional model with the multi-dimensional data attributes, assists in realizing core functions such as virtual inspection, risk control and scheduling decision-making, and improves the intelligent operation level of the station. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0052] Figure 1 The process flow chart of the comprehensive management and control method based on digital twin and multi-source heterogeneous data processing provided by the embodiment of the present invention Figure 1 ;

[0053] Figure 2 The process flow chart of the comprehensive management and control method based on digital twin and multi-source heterogeneous data processing provided by the embodiment of the present invention Figure 2 ;

[0054] Figure 3 This is a structural block diagram of the comprehensive management and control system based on digital twins and multi-source heterogeneous data processing provided by an embodiment of the present invention.

[0055] In the figure: 1. Model conversion module; 2. Optimization module; 3. Fusion module. DETAILED DESCRIPTION

[0056] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0057] Please also refer to Figures 1 to 3 The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing provided by the present invention is now described. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing includes the following steps:

[0058] S1. Scan the site to obtain point cloud data of the surrounding environment, convert the point cloud data into a model, and perform reverse modeling to reconstruct the geometric shape and structure of the building based on the point cloud data to obtain a digital twin model. The digital twin model obtained by reverse modeling is calibrated with the site.

[0059] The surrounding environment includes the geometry and details of the building or site.

[0060] Step S1 preferably includes:

[0061] S1.1. A large amount of point cloud data is obtained through the 3D laser scanning process, which is expressed as:

[0062]

[0063] in, represents a point cloud dataset, Indicates the first points.

[0064] De-noising is performed on the point cloud data. Preferably, noise points are removed by filtering formulas to improve the quality of the data.

[0065] The filtering formula is:

[0066] in, represents the point after filtering, Indicates a point The point set in the neighborhood, Indicates a point The distance to its neighboring points, Indicates the set distance threshold, used to determine whether it is a noise point. Represents the empty set, representing the points removed in the filter.

[0067] During 3D laser scanning, the scanning device may obtain point cloud data from multiple angles or positions under different perspectives. The results of each scan are usually partial and incomplete. Different point cloud data need to be accurately aligned through rigid body transformation operations such as rotation and translation to obtain 3D point cloud data of the complete scene.

[0068] S1.2. Use the ICP (Iterative Closest Point) algorithm for point cloud registration. The formula is:

[0069]

[0070] Among them, T represents the rigid body transformation matrix including the rotation matrix and displacement vector, which is used to align the point cloud data. express The new coordinate position after the transformation matrix T. and represents corresponding points from two different point clouds, represents the Euclidean distance metric.

[0071] S1.3. Further generate a three-dimensional model, i.e., a digital twin model, from the point cloud data through surface fitting. The least squares method is used for fitting:

[0072]

[0073] in, represents the function of fitting the surface, are the surface parameters to be solved, represents a point in the point cloud, represents the fitting error.

[0074] The goal of surface fitting is to generate a smooth 3D model from a point cloud. The accuracy and consistency of the point cloud data directly impact the fitting results. Point cloud registration is a prerequisite for surface fitting. Registered point cloud data is more complete and accurate, which improves the accuracy of surface fitting and ensures the precision and quality of the fitted model.

[0075] S1.4. Compare the model with the actual environment data to verify consistency; the comparison formula is:

[0076]

[0077] in, represents the consistency measure between the digital twin and the actual environment, represents the calibration matrix of the adjustment parameters, represents the model points generated by reverse modeling, Represents actual data points from a production environment.

[0078] In this step, 3D laser scanning and reverse modeling are used to scan the site, capture the geometry and details of the building or site, and capture the exterior and interior structure. Point cloud processing is used to filter, register, and segment the point cloud data to improve data quality and accuracy. The point cloud data is converted into a model, enabling reverse modeling to reconstruct the building's geometry and structure based on the point cloud data. Reverse modeling reconstructs the building's geometry through surface fitting. The reverse model is then calibrated against the site to ensure consistency between the digital twin environment and the on-site production environment, thereby supporting solution verification within the digital twin environment.

[0079] S2. Optimize the digital twin model to improve model performance and enhance the user experience. This will resolve the issue of loading large models and enable high-performance and smooth scene display.

[0080] Step S2 preferably includes:

[0081] S2.1. Reduce the surface area of ​​the digital twin model while ensuring the visual precision of the digital twin model;

[0082] S2.2. Use LOD (Level of Detail) technology to represent the same model with models of different levels of detail. The use of LOD technology can significantly reduce the number of triangles that need to be drawn in a 3D scene without changing the realism of the scene.

[0083] Merge identical components within a model. For identical components within a model, only a single component model is stored, and the remaining identical component models are copied and added to the scene, establishing model reuse relationships and reducing memory usage. This significantly saves system resources while ensuring excellent visualization effects.

[0084] LOD technology adjusts the level of detail based on the distance from the model to the observer:

[0085]

[0086] in, Indicates the LOD level, Represents the distance from the model to the observer, represents the resolution factor, Indicates the proportion of the model on the screen.

[0087] Calculate the rendering complexity of the model:

[0088]

[0089] in, Indicates rendering complexity, Indicates the number of triangles in the model, Indicates the amount of rendering calculation for each triangle.

[0090] Finally, calculate the scene rendering time after LOD technology:

[0091]

[0092] in, Indicates the total rendering time, Indicates the number of models in the scene, Indicates the The basic rendering time of each model, Indicates the The LOD level of the model.

[0093] Example 1: Taking a single station model operation as an example, the scene includes: station main body model, equipment model and component model. The specific model information is:

[0094] Station model: high-precision model (100,000 triangles, LOD=1), medium-precision model (50,000 triangles, LOD=2), low-precision model (10,000 triangles, LOD=3). Rendering time for the high-precision model is 10 seconds, corresponding to the computational complexity parameters of the triangles. In the actual rendering process, the F parameter is often estimated through experiments or experience.

[0095] Device models: high-precision model (50,000 triangles, LOD=1), medium-precision model (25,000 triangles, LOD=2), low-precision model (5,000 triangles, LOD=3). Rendering with the high-precision model takes 5 seconds, corresponding to the computational complexity parameters of the triangles. .

[0096] Component models: high-precision model (200,000 triangles, LOD=1), medium-precision model (100,000 triangles, LOD=2), low-precision model (20,000 triangles, LOD=3). Rendering time for the high-precision model is 3 seconds, corresponding to the computational complexity parameters of the triangles. .

[0097] All high-precision models are used for rendering. According to the above scene rendering time formula, the total rendering time is:

[0098]

[0099] Based on the distance between the viewing angle and each building, the corresponding level of detail is selected using the LOD formula:

[0100] The station model is 100 meters away and uses a high-precision model, corresponding to LOD=1;

[0101] The device model is 200 meters away, using a medium-precision model, corresponding to LOD=2;

[0102] The component model is 300 meters away and uses a low-precision model, corresponding to LOD=3.

[0103] The calculation complexity of each model is calculated according to the rendering complexity formula of the calculation model:

[0104]

[0105]

[0106]

[0107] By reasonably selecting the total rendering time after optimizing the models at different LOD levels, the total rendering time of the scene is calculated according to the formula:

[0108]

[0109] LOD technology significantly improves the rendering performance of three-dimensional scenes by reducing the rendering complexity of distant objects, reducing the rendering time from 18 seconds when using high-precision models to 13.5 seconds. This reduces system resource usage and improves the efficiency of real-time rendering while maintaining visual quality.

[0110] S2.3. Bake the detailed model with a low-poly model to generate models of different precisions. When used on the platform, the model with different precisions is displayed according to the distance between the viewpoint and the model. That is, the low-poly model is used when the model is far away, and the high-poly model is used when the model is close.

[0111] S3. Adopt multi-source heterogeneous data fusion technology to process and classify heterogeneous data from various rail transit disciplines, achieving standardized data processing. This will solve the problem of efficient and rapid application of data on the smart operation and maintenance platform.

[0112] Step S3 preferably includes:

[0113] S3.1. Standardize heterogeneous data; that is, convert data in different formats, such as device status data and sensor data, into a unified representation. The data standardization formula is:

[0114]

[0115] in, Represents the original data value, represents the mean of the data, Represents the standard deviation of the data. By normalizing the data to a uniform scale, it ensures that the data is compared and processed on the same dimension.

[0116] S3.2 Multi-source data fusion processing: This involves mapping, associating, and merging data from different systems through fusion algorithms to ultimately form unified structured data.

[0117] The fusion algorithms include:

[0118] Key-value-based association matching algorithm. If the same primary key (such as equipment number, station ID) exists in two sets of data sources, data will be merged through key-value matching:

[0119]

[0120] in, is the fused dataset, 、 are two data sets to be fused, Represents a primary key-based association operation.

[0121] Semantic-based heterogeneous data matching algorithm. For heterogeneous data sources without clear key values, a semantic matching algorithm is used to determine the correspondence between different data sources by comparing the semantic similarity of fields or data:

[0122]

[0123] in, Represents the fields in data sources A and B, and The similarity between The similarity of attributes, Represents the total number of attributes in the data source.

[0124] In this step, heterogeneous data standardization converts data in different formats such as SQL, JSON, and XML into a unified table structure, associates device status with sensor data through device ID (primary key), and uses semantic matching algorithms to semantically match device descriptions with data such as device status to achieve multi-source data fusion processing.

[0125] S3.3. Streamline the collection, classification, and distribution of heterogeneous data across rail transit disciplines, unifying the mapping of three different coding systems: equipment, asset, and component. This will address the pain point of asset management, where multiple codes exist for one object, and achieve precise mapping between digital and physical space models.

[0126] This step may include:

[0127] S3.3.1. Code standardization. The code needs to be cleaned and formatted to extract the valid part from the code string. Use regular expression parsing to extract the matching string:

[0128]

[0129] in, is the original encoding, is a function that matches regular expressions. This is a normalized code. For example, if the device code is "DEV-12345," the asset code is "ASSET-54321," and the component code is "COMP-67890," regular expression matching and normalization are used to obtain the numeric portion of the device code as the device ID: deviceid = "12345," assetid = "54321," and componentid = "67890."

[0130] This step can effectively solve the problem of inconsistent formats of equipment codes, asset codes and component codes extracted from different systems or databases.

[0131] S3.3.2. Standardize the three different codes: device code, asset code, and component code. In this step, a unique code is generated by string concatenation.

[0132]

[0133] in, 、 、 Represents equipment code, asset code and component code respectively. Represents a string concatenation operation.

[0134] For example, after normalization, , , , the result after splicing is .

[0135] S3.3.3. Perform hash mapping on the code. That is, further use the hash algorithm to generate a unique identifier of fixed length:

[0136]

[0137] in, Represents the concatenated string, represents a hash function, Represents a fixed-length code generated by a hash algorithm.

[0138] This step ensures that the generated code is unique and of fixed length, making it easy to store and retrieve. It effectively solves the problem of long code lengths in actual processing, which leads to slow retrieval and storage efficiency.

[0139] The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing provided by the present invention, compared with the existing technology, realizes efficient data processing and spatial topological structure construction, realizes lightweight and real-time rendering of large models, improves the smooth loading and display of large scenes and large-volume models, and ensures high-performance work in big data scenarios. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing performs standardized processing on data from different systems, realizes cross-system data sharing and calling, facilitates system data intercommunication, and improves data utilization efficiency. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing realizes the precise mapping of the digital twin space to the real station environment, can associate the three-dimensional model with the multi-dimensional data attributes, assists in realizing core functions such as virtual inspection, risk control and scheduling decision-making, and improves the intelligent operation level of the station.

[0140] The present invention also provides a comprehensive management and control system based on digital twins and multi-source heterogeneous data processing. Please refer to Figures 1 to 3 , used to execute the comprehensive management and control method based on digital twins and multi-source heterogeneous data processing as described in any of the above items, including: a model conversion module 1, an optimization module 2 and a fusion module 3, the model conversion module 1 is used to scan the site, obtain point cloud data of the surrounding environment, reconstruct the geometric shape and structure of the building according to the point cloud data, and calibrate the model obtained by the reverse modeling with the site; the optimization module 2 is used to optimize the model; the fusion module 3 is used to adopt multi-source heterogeneous data fusion technology to process and classify heterogeneous data of various professions in rail transit, so as to realize standardized processing of data.

[0141] The present invention provides a comprehensive management and control system based on digital twins and multi-source heterogeneous data processing, constructs a centrally controlled, controllable and manageable three-dimensional intelligent operation and maintenance management platform, and uses the urban rail model as a carrier to solve the problems of dispersion and fragmentation of various professional information systems in the urban rail industry and the inability to share dynamic information. It provides functional services in multiple fields such as assets, space, intelligence, safety, and energy saving, and realizes the reliability of fault prediction and the scalability of the system.

[0142] This invention provides a comprehensive management and control method and system based on digital twins and multi-source heterogeneous data processing. By applying digital twin technology and multi-source heterogeneous data fusion, specialized subsystems are unified and integrated, developing functional modules that meet the needs of each discipline. This provides a unified platform for subway operations managers, technicians, and production personnel, achieving an intelligent upgrade to traditional business needs. These functional modules include, but are not limited to, fault alarm modules, virtual inspection modules, passenger flow monitoring modules, and risk emergency response modules. Leveraging data carriers for multiple scenarios and applications in rail transit operations and maintenance, processes such as model mapping and equipment and facility operational data correlation are accomplished, enabling enterprise-level intelligent operations and maintenance management.

[0143] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A comprehensive management and control method based on digital twins and multi-source heterogeneous data processing, characterized in that: The following steps are involved: Scan the site to obtain point cloud data of the surrounding environment, reconstruct the building's geometry and reverse modeling of the structure based on the point cloud data, and calibrate the model obtained by reverse modeling with the site; Optimize the model; Adopt multi-source heterogeneous data fusion technology to process and classify heterogeneous data of various rail transit disciplines, and realize standardized data processing; The method includes scanning the site to obtain point cloud data of the surrounding environment, reconstructing the geometric shape and structure of the building according to the point cloud data, and calibrating the model obtained by the reverse modeling with the site, including: Obtain a large amount of point cloud data through the 3D laser scanning process, perform denoising on the point cloud data, and perform point cloud registration operations; The point cloud data is further converted into a three-dimensional model through surface fitting; the fitting formula is: Among them, f(c,p i ) represents the function of fitting the surface, c is the surface parameter to be solved, p i represents a point in the point cloud, ‖·‖ represents the fitting error; Compare the model with the data in the actual environment to verify the consistency; the comparison formula is: Among them, D s represents the consistency measure between the digital twin and the actual environment, T represents the calibration matrix of the adjustment parameters, and M(p i ) represents the model point generated by reverse modeling, R(p i ) represent actual data points from the production environment.

2. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 1, characterized in that: The method of obtaining a large amount of point cloud data through the three-dimensional laser scanning process, denoising the point cloud data, and performing point cloud registration operations includes: The registration formula used is: Where T represents the rigid body transformation matrix including the rotation matrix and the displacement vector, p i and q i represents corresponding points from two different point clouds, ‖·‖ represents the Euclidean distance metric, T(p i ) indicates p i The new coordinate position after the transformation matrix T.

3. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 2, characterized in that: The optimization process of the model includes: using models of different degrees of refinement to represent the same model; wherein the level of detail is adjusted according to the distance from the model to the observer using LOD technology: Among them, L represents the LOD level, d represents the distance between the model and the observer, r represents the resolution factor, and s represents the proportion of the model on the screen; Calculate the rendering complexity of the model: C=T×F Among them, C represents the rendering complexity, T represents the number of triangles in the model, and F represents the rendering calculation amount of each triangle; Calculate the scene rendering time after LOD technology: Among them, T r represents the total rendering time, n represents the number of models in the scene, T i Indicates the basic rendering time of the i-th model, LOD i Indicates the LOD level of the i-th model.

4. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 1 or 3, characterized in that: The optimization process of the model includes: performing low-poly baking on the fine model to generate models of different precisions, wherein when in use, the selection of the fineness of the model depends on the distance between the viewpoint and the model.

5. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 1, characterized in that: The optimization process of the model includes: Standardize heterogeneous data; the data standardization formula is: Among them, x represents the original data value, μ represents the mean of the data, and σ represents the standard deviation of the data; by standardizing the data to a unified scale, it ensures that the data is compared and processed in the same dimension; Through the fusion algorithm, the data of different systems are mapped, associated and merged to form unified structured data; Carry out process-based collection, classification and distribution of heterogeneous data of various professions in rail transit, unify the mapping of three different coding systems of equipment coding, asset coding and component coding, and generate a unique code.

6. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 5, characterized in that: The data of different systems are mapped, associated and merged through the fusion algorithm to finally form unified structured data. The fusion algorithm includes: one or more of a key-value-based association matching algorithm and a semantic-based heterogeneous data matching algorithm.

7. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 5, characterized in that: The process of collecting, classifying and distributing heterogeneous data of various rail transit disciplines is described, and the three different coding systems of equipment coding, asset coding and component coding are uniformly mapped to generate a unique code, including: Clean and format the code to extract the valid part from the code string; The three different codes of equipment code, asset code and component code are standardized; Hash the encoding to generate a unique identifier of fixed length.

8. The comprehensive management and control method based on digital twins and multi-source heterogeneous data processing according to claim 7, characterized in that: The encoding cleaning and formatting is used to extract the valid part from the encoded string, including: using regular expression parsing to extract the matching string: C standard =regex_match(C raw ) Among them, C raw is the original code, regex_match is the regular expression matching function, C standard It is the normalized encoding.

9. Comprehensive management and control system based on digital twins and multi-source heterogeneous data processing, characterized by: A comprehensive management and control method based on digital twins and multi-source heterogeneous data processing as described in any one of claims 1 to 8, comprising: The model conversion module is used to scan the site, obtain point cloud data of the surrounding environment, reconstruct the geometric shape and structure of the building based on the point cloud data, and calibrate the model obtained by reverse modeling with the site; Optimization module, used to optimize the model; The fusion module is used to collect and classify heterogeneous data from various rail transit disciplines in a process-based manner using multi-source heterogeneous data fusion technology to achieve standardized data processing; The method includes scanning the site to obtain point cloud data of the surrounding environment, reconstructing the geometric shape and structure of the building according to the point cloud data, and calibrating the model obtained by the reverse modeling with the site, including: Obtain a large amount of point cloud data through the 3D laser scanning process, perform denoising on the point cloud data, and perform point cloud registration operations; The point cloud data is further converted into a three-dimensional model through surface fitting; the fitting formula is: Among them, f(c,p i ) represents the function of fitting the surface, c is the surface parameter to be solved, p i represents a point in the point cloud, ‖·‖ represents the fitting error; Compare the model with the data in the actual environment to verify the consistency; the comparison formula is: Among them, D s represents the consistency measure between the digital twin and the actual environment, T represents the calibration matrix of the adjustment parameters, and M(p i ) represents the model point generated by reverse modeling, R(p i ) represent actual data points from the production environment.

Images