Highway geological disaster risk early warning method and system based on BIM and numerical simulation

Abstract

The present invention is a highway geological disaster risk warning method and system based on BIM and numerical simulation. It mainly combines drone close photogrammetry technology and BIM technology to establish a three-dimensional geological model of the highway and the geological environment along it, and uses ABAQUS software and PFC discrete element simulation to simulate different working conditions to perform high-speed analysis of the scope of geological disaster damage under external conditions. The integration of drone mapping, BIM technology, numerical simulation (ABAQUS and PFC) and GIS platform (ArcGIS) realizes accurate and dynamic risk warning of in-service highways. The present invention can realize multi-level geological disaster risk warning from static structure to dynamic environment, analyze the damage range and movement process of geological disasters, and display the relationship between the scope of geological disaster impact and highway in the real-life model, especially in terms of slope landslide, roadbed settlement, etc. It has significant innovative advantages.

Description

Technical Field

[0001] This invention relates to the fields of engineering geology and transportation engineering research, and more specifically, to a highway geological hazard risk warning method and system based on BIM and numerical simulation. This method combines drone close-range photogrammetry with BIM and numerical simulation technology to provide risk warnings for geological hazards on in-service highways. Background Art

[0002] With the rapid development of transportation infrastructure, expressways have become a vital component of the national economy. However, these highways, exposed to complex natural environments for extended periods, are susceptible to multiple factors, including geological disasters, climate change, and traffic volume, posing risks to their safety and stability. Therefore, real-time, accurate risk monitoring and early warning for in-service expressways have become crucial for ensuring their safe operation.

[0003] In recent years, drone-based photogrammetry technology has made significant progress in the field of geographic information collection. Equipped with high-precision cameras and GPS systems, drones can rapidly acquire high-quality image data and generate 3D models of target areas using 3D modeling techniques. Furthermore, Building Information Modeling (BIM), an advanced engineering design and management tool, has been widely adopted in highway engineering design and road geological surveys.

[0004] When it comes to surveying and modeling existing highways, both drone oblique photography and BIM technology currently have their own limitations. Given that highways often reside in complex terrain, such as mountainous and hilly areas, most studies rely solely on drone oblique photography for modeling. However, due to the large terrain undulations or dense vegetation, drone oblique photography can be less accurate, and flight altitude restrictions prevent full coverage of the entire survey area. For example, in "Research on Highway Slope Data Acquisition Models Based on Unmanned Aerial Vehicle Systems," Yang Yanwei used drones to collect highway slope data, but due to weak signals, the data obtained was fuzzy, and the subsequent data recovery was also inaccurate. Furthermore, existing technologies do not apply to geological risk warnings for highways and surrounding facilities, focusing solely on modeling the highways themselves.

[0005] When dealing with complex terrain, traditional methods suffer from limitations such as long data collection cycles, low accuracy, and poor real-time performance. These limitations make large-scale, dynamic, and real-time monitoring difficult to achieve, and they are easily affected by external factors such as climate and topography. Existing methods often rely on fixed monitoring points, making them inflexible and unable to adapt to terrain changes and emergencies. Therefore, when highways are located in areas at high risk of geological disasters, existing methods are unlikely to provide effective early warnings. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a highway geological hazard risk warning method and system based on BIM and numerical simulation. This method aims to address the current problems of difficult and inaccurate highway terrain data acquisition, mismatch between designed and actual terrain environments, and delayed risk warnings. This method enhances the accuracy and reliability of highway risk warnings and is of great significance for the subsequent use of numerical simulation techniques to calculate the stability and range of motion of geological bodies surrounding highways under rainfall conditions.

[0007] The technical solution adopted by the present invention to solve the technical problem is:

[0008] In a first aspect, the present invention provides a highway geological disaster risk early warning method based on BIM and numerical simulation, the steps of the early warning method are:

[0009] Step 1: Use drone close-up photogrammetry to obtain high-precision original point cloud data and image data along the target highway and surrounding areas;

[0010] Step 2: Import the raw point cloud data collected by the drone into Context Capture software. ContextCapture then performs denoising, optimization, and precision adjustment on the raw point cloud data, removes invalid and duplicate points, and stitches and integrates data from different flight batches to generate a complete 3D point cloud model. The spatial coordinate system is then adjusted according to project requirements to ensure consistency with the coordinates of the subsequently established BIM model, completing the preprocessing of the point cloud data.

[0011] Step 3: Based on the data pre-processed in step 2, Smart 3D software is used to generate high-precision real-life 3D terrain of the in-service road and surrounding areas, and the real-life 3D terrain data is converted into a data format;

[0012] Step 4: Import the point cloud data pre-processed in Step 2 into Revit software for 3D BIM modeling. Use Revit's terrain modeling tools to generate a terrain surface that accurately reflects the actual terrain features of the highway and its surrounding areas. Based on the existing design data of the highway infrastructure, use Revit to create a highway infrastructure model. At the same time, combine the terrain features in the point cloud data (such as trees, gullies, etc.) to refine the model details and avoid inaccurate modeling due to occlusion or errors. Through comparative analysis of the terrain features in the point cloud data and the infrastructure model, verify and optimize the size and position of the BIM model to ensure that it fully matches the actual terrain and facility environment, thereby obtaining a BIM model of the in-service road and surrounding houses based on the point cloud data.

[0013] Step 5: Compare and adjust the BIM model with the high-precision real-world 3D terrain generated by Smart3D software to ensure seamless integration of the facilities in the BIM model and the terrain in the real-world 3D terrain. Then, integrate the BIM model with the high-precision real-world 3D terrain generated by Smart3D software and optimize the geometric details of the BIM model to ensure that the BIM model meets design requirements and truly reflects actual terrain conditions. In Revit, export the BIM model that meets the requirements into a format suitable for processing in ArcGIS software.

[0014] Step 6: Export the BIM model processed in step 5 from Revit to STL format or OBJ format for ABAQUS numerical simulation and analysis. The STL format data is used for three-dimensional numerical simulation modeling and analysis, and the OBJ format is used for two-dimensional section numerical simulation modeling and analysis. Export the DXF format from Revit to PFC numerical simulation and analysis.

[0015] For two-dimensional modeling and analysis, using OBJ format data, we first establish a complete slope geometry model using ABAQUS software for finite element analysis. The physical properties and boundary conditions of the slope are set, and the slope stress and displacement fields under rainfall conditions are simulated. When the convergence conditions are reached, the area of potential instability, i.e., the instability range, is identified.

[0016] After importing the DXF file exported from Revit into PFC, a 2D cross-sectional model of the landslide was generated. The sliding mass and sliding bed of the slope were divided according to the instability range under rainfall conditions determined by ABAQUS. Discrete element analysis was performed on the potentially unstable area (i.e., the sliding mass) using PFC software to dynamically simulate the movement of particles during the landslide. Dynamic conditions (such as rainfall and earthquakes) were applied to simulate the sliding path, accumulation range, and displacement of particles, as well as their kinematic characteristics under earthquakes.

[0017] For three-dimensional modeling and analysis, using STL format data, we first used ABAQUS software to build a complete slope geometry model for finite element analysis. The physical properties and boundary conditions of the slope were set, and the slope stress and displacement fields under rainfall conditions were simulated. When the convergence conditions were reached, the areas that might be unstable were identified.

[0018] The DXF file exported from Revit was imported into PFC to generate a 3D model of the landslide. The sliding mass and sliding bed of the slope were divided according to the instability range under rainfall conditions determined by ABAQUS. Discrete element analysis was performed on the potentially unstable area (i.e., the sliding mass) using PFC software to dynamically simulate the movement of particles during the landslide. Dynamic conditions (such as rainfall and earthquakes) were applied to simulate the sliding path, accumulation range, and displacement of particles, as well as their kinematic characteristics under earthquakes.

[0019] Step 7: Export the numerical simulation results of PFC into a format suitable for processing by ArcGIS software and import it into the ArcGIS platform. Simultaneously, import the BIM model from step 5 into the ArcGIS platform. Set the BIM model and the landslide data imported from PFC to share the same coordinate system so that they can be correctly displayed and three-dimensionally superimposed within the same spatial range.

[0020] Step 8: Use ArcGIS analysis tools to analyze the relative relationship between the sliding path of landslide particles, the accumulation area, and the facilities in the BIM model, and intuitively display the interaction between the landslide area, accumulation location, and road facilities. Ultimately, form early warning analysis and maintenance recommendations, and generate a visual risk warning map.

[0021] Furthermore, the method can provide dynamic early warning. The data collected by the drone communicates with the Context Capture software, the Context Capture software communicates with the Smart 3D software and the Revit software respectively, the Revit software communicates with the ABAQUS software and the PFC software respectively, and at the same time, the ABAQUS software communicates with the PFC software, and the Revit software and the PFC software both communicate with the ArcGIS platform. The drone dynamically collects data, so that the BIM model can be continuously updated according to the periodic data collection of the drone, and the early warning adjustment is performed in combination with the simulation results of the PFC, thereby maintaining the dynamic nature and accuracy of the BIM model and ensuring that the early warning information always conforms to the current environment.

[0022] The ABAQUS software also communicates with the ArcGIS platform to display the numerical simulation results of the ABAQUS software.

[0023] In step 5, the facilities in the BIM model include roads, bridges, tunnels and surrounding mountains.

[0024] The process of processing point cloud data with the help of Smart 3D software includes: matching points with the same name using a feature-based matching algorithm, performing multi-view image joint adjustment using the bundle method regional block adjustment algorithm, and completing the preliminary construction of a high-density point cloud based on the clustering algorithm and the patch-based dense matching algorithm; finally, the TIN network model constructed based on the point cloud is combined with texture mapping to achieve the construction of a high-precision real-life 3D model.

[0025] In a second aspect, the present invention provides a highway geological disaster risk early warning system based on BIM and numerical simulation, the system comprising:

[0026] The UAV data acquisition module uses drones to collect high-precision raw point cloud data and image data along the target highway and surrounding areas;

[0027] A data preprocessing module is used to preprocess the data collected by the UAV data acquisition module;

[0028] The BIM modeling module is used to perform BIM modeling based on pre-processed data and optimize the BIM model using the real-life 3D terrain generated by Smart 3D software, thereby obtaining a BIM model that meets design requirements and truly reflects actual terrain conditions.

[0029] The numerical simulation module is used to perform finite element analysis and discrete element analysis based on the BIM model output by the BIM modeling module. The finite element analysis is used to determine the area of possible instability when the convergence conditions are reached. The discrete element analysis is used to determine the sliding path, accumulation range and displacement of particles in the area of possible instability, as well as the kinematic characteristics under earthquake action.

[0030] The GIS data integration module is used to integrate and display the output results of the BIM modeling module and the numerical simulation module, analyze the relative relationship between the sliding path of landslide particles, the accumulation area and the facilities in the BIM model, and intuitively display the interaction between the landslide area, accumulation location and road facilities;

[0031] The risk warning and feedback module is used to conduct risk assessment based on the output of the GIS data integration module, and generate a visual risk warning map based on the final warning analysis and maintenance recommendations.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] 1. To address the current problem of insufficient early warning of geological disaster risks along in-service expressways, this paper mainly combines drone-based photogrammetry technology with BIM technology to establish a three-dimensional geological model of the expressway and the geological environment along it. ABAQUS software and PFC discrete element simulation are used to simulate different working conditions to conduct high-speed analysis of the scope of geological disaster damage under external conditions.

[0034] 2. This invention effectively integrates drone close-up photogrammetry, BIM technology, numerical simulation tools (such as ABAQUS and PFC), and GIS. Using drone close-up photography, it acquires high-precision terrain data along highways in real time, generating high-precision BIM models to meet the geological monitoring needs of complex terrain and achieve all-round dynamic monitoring of highway slopes. Combined with simulation analysis using ABAQUS and PFC, this invention can achieve multi-level geological disaster risk warnings from static structures to dynamic environments, analyze the damage range and movement process of geological disasters, and display the relationship between the geological disaster impact range and highways in a real-life model. It has significant innovative advantages in terms of slope landslides and roadbed settlement.

[0035] 3. UAV-based photogrammetry technology, through efficient and accurate terrain data collection, can quickly generate 3D models of highways. Integrating these aerial survey-based models with BIM further enhances the model's geometric accuracy and information richness, ensuring the design perfectly matches the actual terrain. This combination reduces manual modeling errors, improves efficiency, and enhances data accuracy.

[0036] 4. Combining Smart 3D software with BIM not only accurately displays terrain details but also integrates information such as road project design, construction, and operation and maintenance onto a unified visualization platform. This integrated display facilitates multi-dimensional monitoring and analysis of highway safety. High-precision raw point cloud and image data acquired through drone photogrammetry provides accurate three-dimensional terrain support for BIM models. Combined with BIM technology's full lifecycle data management capabilities, monitoring information can be fed back to the BIM system in real time. Dynamic, real-time risk warnings are achieved through numerical simulations using ABAQUS and PFC. ArcGIS overlays and analyzes BIM models, stress fields, and particle motion to form a multi-dimensional early warning model, providing data support for highway lifecycle management.

[0037] 5. During the numerical simulation phase, the terrain portion of the BIM model is exported to a format suitable for ABAQUS and PFC (such as STL and DXF). ABAQUS simulates the slope stress and displacement fields under different rainfall conditions through finite element analysis, while PFC generates a particle model based on the BIM model profile for landslide particle flow analysis and dynamic simulation of the landslide movement process. This combination integrates the static geometric model of BIM with the dynamic analysis of numerical simulation, realizing dynamic monitoring of slopes under complex geological conditions. This improves the risk prediction capability of the highway during its in-service phase. This data linkage and real-time warning helps to detect potential road safety hazards early, take timely measures, and ensure the safe operation of the highway.

[0038] 6. Flexibility and adaptability: The present invention still has efficient data collection capabilities in complex terrains, is suitable for highway monitoring in different geological environments, and provides higher adaptability for disaster warning.

[0039] 7. Improved costs and efficiency: Drones can quickly collect data over a large area, reducing labor and time costs, improving monitoring efficiency, and enabling early warning in large areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 : Numerical simulation analysis diagram of geological hazards along in-service expressways.

[0041] Figure 2: Schematic diagram of risk early warning for highways affected by geological disasters in the method of the present invention AcrGIS.

[0042] Figure 3 : Schematic diagram of traditional methods for early warning of highways and houses along the highway.

[0043] Figure 4 : Schematic diagram of the process of the risk early warning system for in-service highways of the present invention. DETAILED DESCRIPTION

[0044] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. Of course, the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0045] In this invention, drones and BIM are primarily used for modeling and extracting terrain information, while ABAQUS and PFC are used to analyze the impact range of geological hazards along the route. By using the distance and speed between geological hazards and highways in the BIM model, early warnings of geological hazard risks in different highway sections are issued, and the warning levels can be accurately distinguished. Existing technologies can only roughly determine the impact range of geological hazards, roughly assign warning levels, and issue warning information. This application significantly improves the accuracy of existing technologies.

[0046] This invention innovatively integrates drone mapping, BIM technology, numerical simulation (ABAQUS and PFC), and a GIS platform (ArcGIS) to achieve accurate and dynamic risk warnings for in-service highways. The following are the organic combinations of these software in this application:

[0047] Combining drone mapping with BIM model creation

[0048] Drones collect high-precision raw point cloud data and images of the highway and surrounding terrain to generate a basic terrain model. After pre-processing, such as filtering and denoising, the raw point cloud data is converted into processed point cloud data, which can be directly imported into BIM software (such as Revit or Navisworks) to create 3D models of terrain, roads, bridges, and other structures. The BIM model combines road design with field drone measurements to ensure a high degree of accuracy and accurately matches the actual terrain, laying a precise data foundation for subsequent numerical simulations.

[0049] Combination of BIM model with ABAQUS and PFC

[0050] During the numerical simulation phase, the terrain portion of the BIM model was exported to formats suitable for ABAQUS and PFC (such as STL and DXF). ABAQUS simulated the slope stress and displacement fields under various rainfall conditions through finite element analysis. PFC, based on the ABAQUS analysis results and BIM model data, generated a particle model for landslide particle flow analysis and dynamic simulation of landslide movement. This combined approach merged the static geometric model of BIM with the dynamic analysis of numerical simulation, enabling dynamic monitoring of slopes under complex geological conditions.

[0051] Combination of numerical simulation and GIS platform

[0052] After completing the numerical simulation, the analysis results from ABAQUS and PFC (such as stress, displacement, and particle motion paths) can be imported into the ArcGIS platform and overlaid with the BIM model for 3D visualization. ArcGIS compares the spatial relationship between the simulation results and road infrastructure, visually displaying the interaction between landslide areas, accumulation locations, and road infrastructure, ultimately generating early warning analysis and maintenance recommendations. The GIS platform provides a global view, helping decision makers dynamically understand the spatial distribution of landslide risk.

[0053] Data format and compatibility issues

[0054] Data formats vary significantly between different software, such as the LAS format for drone raw point cloud data, the IFC format for BIM software, the STL format for ABAQUS, and the DXF format for PFC. This application addresses compatibility issues between these formats through data format conversion. For example, by converting raw point clouds to OBJ or STL formats, this data can be seamlessly integrated into BIM modeling and finite element analysis.

[0055] Data accuracy and real-time challenges

[0056] Traditional risk warning models struggle to provide high-precision, real-time risk warnings in complex terrain. This invention leverages high-precision drone data sources to collect real-time road and surrounding terrain data through close-up photogrammetry, generating a detailed BIM model and combining it with numerical simulation for dynamic analysis. In particular, PFC's simulation of the flow of landslide particles provides real-time insights into the impact of terrain changes on highways, providing a reliable basis for real-time warnings in complex terrain.

[0057] Integration and display of multidimensional data

[0058] The discrepancies between multidimensional data, such as terrain data collected by drones, BIM models, ABAQUS stress fields, and PFC landslide particle flow data, were overlaid on the ArcGIS platform using WS, enabling unified presentation through 3D visualization analysis. This integrated approach intuitively displays the spatial relationship between landslide areas and road infrastructure, empowering risk warning systems with the ability to integrate multidimensional information.

[0059] Dynamic model update and accuracy maintenance

[0060] In practice, the terrain and facilities along highways may change, making it difficult for traditional static models to reflect this dynamic information. This invention enables the BIM model to be continuously updated based on periodic data collection from drones and adjusted based on simulation results from ABAQUS and PFC. This maintains the model's dynamic nature and accuracy, ensuring that warning information remains relevant to the current environment.

[0061] This patented system achieves precise, real-time monitoring of existing highway slopes through a multi-software combination of drones, BIM, numerical simulation, and GIS. Compared to traditional methods, this patented system offers highly automated data collection and processing capabilities, significantly improving monitoring accuracy and real-time performance.

[0062] The system integrates multi-dimensional information in complex terrain, upgrading highway risk warning from static monitoring to dynamic warning. This fully considers the real-time changes of geological hazards such as landslides and subsidence, providing more scientific decision-making support for road management.

[0063] The present invention provides a highway geological disaster risk early warning system based on BIM and numerical simulation, the system comprising:

[0064] The UAV data acquisition module uses drones to collect high-precision raw point cloud data and image data along the target highway and surrounding areas;

[0065] A data preprocessing module is used to preprocess the data collected by the UAV data acquisition module;

[0066] The BIM modeling module is used to perform BIM modeling based on pre-processed data and optimize the BIM model using the real-life 3D terrain generated by Smart 3D software, thereby obtaining a BIM model that meets design requirements and truly reflects actual terrain conditions.

[0067] The numerical simulation module is used to perform finite element analysis and discrete element analysis based on the BIM model output by the BIM modeling module. The finite element analysis is used to determine the area of possible instability when the convergence conditions are reached. The discrete element analysis is used to determine the sliding path, accumulation range and displacement of particles in the area of possible instability, as well as the kinematic characteristics under earthquake action.

[0068] The GIS data integration module is used to integrate and display the output results of the BIM modeling module and the numerical simulation module, analyze the relative relationship between the sliding path of landslide particles, the accumulation area and the facilities in the BIM model, and intuitively display the interaction between the landslide area, accumulation location and road facilities;

[0069] The risk warning and feedback module is used to conduct risk assessment based on the output of the GIS data integration module, and generate a visual risk warning map based on the final warning analysis and maintenance recommendations.

[0070] The risk warning system for in-service highways in this application is a complete dynamic monitoring and warning system, which aims to achieve full-process management from data collection to risk assessment. The various modules are interconnected and closely integrated to form a multi-dimensional and multi-functional dynamic risk assessment system.

[0071] The UAV data acquisition module uses drones equipped with LiDAR and high-resolution cameras to efficiently map the terrain along and around highways. By establishing image control points and GPS equipment to ensure data accuracy, the module collects raw point cloud data (LAS format), high-resolution image data (GeoTIFF format), and digital elevation models (DEMs). The drone's automated flight path planning and rapid coverage capabilities provide a convenient and efficient solution for data collection in complex terrain.

[0072] The BIM modeling module imports raw point cloud data collected by drones into BIM software (such as Revit) to generate a 3D terrain model that closely matches the actual terrain. By integrating design data with measured terrain data, this module constructs a detailed 3D model of the roadbed, bridges, mountain, and drainage system, providing a precise geometric foundation for subsequent numerical simulation analysis. The 3D visualization capabilities of the BIM model intuitively reflect the structural relationship between road facilities and the surrounding terrain.

[0073] In the numerical simulation module, ABAQUS and PFC perform static and dynamic risk analysis, respectively. ABAQUS simulates the slope's stress and displacement fields through finite element analysis, predicting potential landslide instability areas under conditions such as rainfall. PFC dynamically simulates the flow paths and accumulation ranges of landslide particles based on cross-sectional data from the BIM model. By combining these two tools, the system accurately quantifies the impact of landslides on road infrastructure, providing strong technical support for risk warning in complex geological conditions.

[0074] The GIS data integration module imports the aforementioned data into the ArcGIS platform, combining it with BIM models and numerical simulation results to integrate and analyze multidimensional data. Using tools such as 3DAnalyst, the GIS platform overlays BIM models, stress distribution, and particle motion trajectories in three-dimensional space, visualizing the specific impact areas of landslides on highway facilities. The system also generates landslide area distribution maps and risk level indicators, helping managers quickly understand the spatial relationships within landslide areas and improving decision-making efficiency.

[0075] The risk warning and feedback module combines numerical simulation results with real-time monitoring equipment to provide intelligent early warnings. ABAQUS software collects real-time data from displacement monitoring sensors and rain gauges, updates the output of the numerical simulation module, and then displays and updates the risk warning map in real time within the GIS platform. The system can quickly identify high-risk areas and trigger alarm mechanisms. Furthermore, the module generates maintenance recommendations based on simulation results to guide the implementation of disaster prevention measures. The high real-time nature of the early warning information effectively reduces the threat posed by highway disasters to operations.

[0076] The system combines drone mapping, BIM modeling, numerical simulation, and GIS visualization analysis to form a closed loop from data collection to risk assessment. Its high-precision data acquisition, efficient dynamic simulation, and multidimensional data integration capabilities are applicable to diverse scenarios such as highways, bridges, and tunnels. It can provide accurate and efficient disaster risk warning solutions for transportation infrastructure, thereby improving operational safety and scientific management.

[0077] Example 1

[0078] The present invention is based on the highway geological disaster risk early warning method of BIM and numerical simulation, and the specific steps are as follows:

[0079] S1. UAV close-up photogrammetry technology to collect terrain information

[0080] First, drones were used to collect topographic information of the highway and its surrounding areas using close-up photogrammetry. Equipped with high-precision cameras, LiDAR, and GPS, the drones continuously photographed the target area along a pre-set route, acquiring high-precision raw point cloud and image data. The collected data included the topographical features surrounding the road, as well as the precise spatial locations of infrastructure such as bridges and tunnels. This data was then used to generate a 3D raw point cloud model with high-precision textures and true geographic locations.

[0081] Specific operation steps: (1) The flight quality is mainly affected by the route, relative altitude, and overlap. To obtain high-quality aerial images, these parameters need to be set when planning the flight. In this embodiment, the measurement is performed using a DJI Mavic 3E drone to obtain the required number of images and the corresponding Pos point data, image width, and height.

[0082] (2) Image control point data acquisition: The number of control points is determined based on the actual situation of the highway being measured, and the coordinates and elevation values of the control points are measured in the field. Based on the route planning and image data, the regional network method is mainly used for point distribution. According to the control point layout principle, appropriate control points within the measurement area are selected as image control points for solution.

[0083] (3) Accuracy requirements for image control points: The error in the image of plane image control points and horizontal image control points relative to adjacent basic image control points shall not exceed 0.2mm. The error in the elevation of elevation image control points and horizontal image control points shall not exceed 0.1m. The positions of image control points shall be marked on the image with a "crosshair" of appropriate size and described in detail with text. After shooting, the drone photos with Pos data shall be imported into Pix4D software for stitching, and dense original point clouds, aerial triangulation, mesh and texture shall be generated to finally obtain digital orthophoto maps (DOM), digital surface models (DSM) and three-dimensional real scene models.

[0084] S2. Use Context Capture to preprocess the original point cloud data

[0085] In S1, the raw point cloud data collected by the drone is imported into Context Capture software, ensuring that the data format is LAS, LAZ, or PLY. This raw point cloud data contains high-precision information about the highway and its surrounding terrain, and serves as the foundation for subsequent data processing. In Context Capture, the raw point cloud data is denoised, optimized, and precision-adjusted to remove invalid and duplicate points, ensuring data clarity and accuracy. If the drone collects data over multiple flights, the software can stitch together the data from different flight batches to generate a complete 3D point cloud model. The spatial coordinate system is then adjusted according to project requirements to ensure consistency with the coordinates of the subsequently constructed BIM model.

[0086] S3 and Smart 3D software generate 3D terrain models based on pre-processed point cloud data

[0087] To ensure the accuracy and precision of the 3D real-world model, the pre-processed point cloud data obtained from S2 was used to generate a high-precision real-world 3D terrain of the terrain where the in-service road is located using Smart3D software. The real-world 3D terrain was also converted to a different data format. Based on this, the image control points before the aerial survey were compared with the corresponding points after the modeling to complete the accuracy analysis.

[0088] Based on the computer vision 3D reconstruction algorithm, Smart3D software can solve the precise exterior orientation elements by simply combining the Pos data obtained during the flight. It also uses a feature-based matching algorithm to match the same-name points, uses the bundle method regional block adjustment method to perform multi-view image joint adjustment, and completes the preliminary construction of the high-density processed point cloud based on the clustering algorithm and the patch-based dense matching algorithm. Finally, the TIN network model constructed based on the processed point cloud is combined with texture mapping to achieve the construction of high-precision real-life 3D terrain (such as Figure 2 shown).

[0089] Smart3D software generates data in various formats, including s3c, fbx, osgb, dae, obj, and stl. OSGB, obj, and fbx formats are compatible with various modeling software, facilitating later modifications and adapting to various display platforms, enabling rapid online publishing and sharing. This paper utilizes the obj format in conjunction with BIM technology.

[0090] S4. Combine pre-processed point cloud data with BIM software to generate road and house models

[0091] (1) Import the pre-processed point cloud data in S2 into Revit software for 3D BIM modeling. First, create a terrain surface model using the elevation information in the point cloud to accurately reflect the actual terrain features of the highway and its surrounding areas. On this basis, create an infrastructure model of the highway according to the design plan, including structures such as roads, bridges, and tunnels. At the same time, combine the terrain features in the point cloud data (such as trees, valleys, etc.) to refine the model details and avoid inaccurate modeling caused by occlusion or errors. By comparing and analyzing the terrain features in the point cloud data with the infrastructure model, the size and position of the BIM model are verified and optimized to ensure that it fully matches the actual terrain and facility environment.

[0092] (2) Export the BIM model from Revit to OBJ format for further analysis and visualization in other software (such as ArcGIS). Through this step, the pre-processed point cloud data was successfully and efficiently combined with BIM technology, and an accurate 3D real scene and infrastructure design model was constructed, providing accurate and reliable data and technical support for landslide risk warning, disaster assessment, and highway maintenance and management.

[0093] Note that steps S3 and S4 can be performed simultaneously, both based on the preprocessed point cloud data obtained in S2. In S3, smart3D is primarily used to model the terrain surrounding the highway, while in S4, Revit is primarily used to model the highway and surrounding buildings due to its greater accuracy. Because smart3D offers greater accuracy for modeling natural terrain, it was used to model the terrain beyond the highway and buildings.

[0094] Combination of S5, BIM model and real 3D terrain generated by Smart3D

[0095] S3 and S4 respectively obtained high-precision real-life 3D terrain and BIM models of the highway and surrounding buildings. In order to achieve the purpose of risk warning, the two need to be combined on the same platform. The specific steps are as follows:

[0096] (1) Format compatibility and data import:

[0097] Export the real-world 3D terrain generated by Smart3D to OBJ or OSGB format (supporting the storage of high-precision data details), and then import it into Revit as a reference. In Revit, the real-world 3D terrain is displayed as a complete real-world 3D background, which can intuitively show the topographic features of the highway and surrounding area.

[0098] (2) Terrain alignment with BIM model:

[0099] In Revit, coordinate system adjustments are used to spatially align the real-world 3D terrain generated by Smart3D with the BIM model. Both models share the same coordinate reference system to ensure seamless integration of BIM model facilities (such as roads, bridges, and tunnels) with the real-world 3D terrain.

[0100] (3) Model integration and analysis:

[0101] In Revit, the BIM model is integrated with the real-world 3D terrain to create a comprehensive model that encompasses both the infrastructure design and the actual terrain. Comparative analysis can then be used to verify the accuracy of the BIM design model, ensuring that it accurately reflects the location and scale of the facilities within the 3D terrain. For example, by comparing the highway subgrade with the actual slope in the real-world 3D terrain, the model can be checked for elevation differences or offsets, allowing adjustments to be made as needed.

[0102] (4) Detail optimization and integration:

[0103] Using Revit tools, Smart3D's real-world 3D terrain serves as a background reference to further optimize the geometric details of the BIM model. For example, adjustments can be made to road slopes, bridge foundation heights, or tunnel excavation angles based on the actual terrain's undulations. This optimization ensures that the BIM model meets design requirements while accurately reflecting actual terrain conditions.

[0104] S6. Combination and simulation analysis of the established model with ABAQUS and PFC software

[0105] (1) Convert the model obtained in S5 into a 3D geometry format that can be used by ABAQUS and PFC. The OBJ format exported from Revit can be used for ABAQUS numerical simulation and analysis, and the DXF format exported from Revit can be used for PFC numerical simulation and analysis.

[0106] (2) In ABAQUS software, import the optimized BIM model of step S5 and establish a complete slope geometry model. Use the built-in meshing tool of ABAQUS to finely divide the mesh to ensure the calculation accuracy of the slope stress field and displacement field. Then, according to the engineering requirements, set the attribute parameters of the geological material (such as the Mohr-Coulomb model) to simulate the slope response under different external loads, such as rainfall, earthquake, etc. Especially under rainfall conditions, ABAQUS can simulate the impact of pore water pressure on the slope and predict the potential area for landslides. Through ABAQUS calculations, the stress distribution, displacement distribution and plastic zone distribution of the slope under different working conditions can be obtained to help identify areas that may become unstable. When the finite element simulation reaches the convergence condition, the area that may become unstable can be identified. At this time, the simulation field and stress field simulated by ABAQUS software are considered accurate.

[0107] (3) Use PFC software to perform discrete element analysis on areas that may become unstable in order to dynamically simulate the movement of particles during the landslide. The discrete element analysis here is performed under the premise that the simulation field and stress field of ABAQUS are accurate. After importing the DXF format exported from Revit into PFC, a two-dimensional cross-sectional model of the landslide is generated. The particles are filled by a random generation method, and the sliding area and sliding bed are set to generate a particle model with a certain porosity. Subsequently, rainfall conditions are applied to the model, the pore water pressure boundary is set, and the movement path, velocity and accumulation position of the landslide particles are tracked by PFC. In addition, PFC can also simulate the impact of earthquakes on slope stability by applying acceleration conditions, thereby analyzing the kinematic characteristics of landslide particles under earthquake action.

[0108] Importing S7 and PFC 2D models and BIM models into ArcGIS software

[0109] (1) PFC model data export: The PFC model is mainly a landslide or geological body model composed of particles. These particles and their movement trajectories are the core of the PFC model. Export the PFC model to a format suitable for GIS software processing. PFC can use export tools to convert PFC simulation results (such as particle position, movement trajectory, etc.) into DXF, STL or CSV formats. STL is suitable for the visualization of three-dimensional models, while CSV can store particle position information for point data visualization. Use the wallexport and geometry export commands to export the landslide boundary and particle position into DXF or STL files, and export the spatial position information of the landslide particles into CSV format (including the X, Y, and Z coordinates of the particles), which is suitable for the visualization of three-dimensional spatial point data in ArcGIS.

[0110] (2) BIM model data export: The BIM model primarily includes the highway and its surrounding buildings and terrain. The BIM model is exported from Revit software into common 3D model formats, including IFC, OBJ, and FBX. These formats can preserve complex geometric structures and are suitable for 3D model comparison and overlay in ArcGIS. Before exporting, the BIM model is simplified, retaining structural information such as roads, bridges, and tunnels, and removing irrelevant details to improve processing efficiency.

[0111] (3) Importing PFC and BIM models into ArcGIS:

[0112] 3.1) Import the PFC model. Use the Add Data function of ArcGIS to import the STL or DXF file exported by PFC into ArcGIS. The STL file can be used to represent the three-dimensional position of the landslide particles, while the DXF file is used to represent the landslide boundary. Then import the CSV file to import the spatial position information of the PFC particles as point data. Use the AddXY Data function of ArcGIS to convert the coordinate data in the CSV file into point features. Set the coordinate system (such as the UTM coordinate system) to ensure that the point data is displayed correctly in three-dimensional space. These points represent the position and movement trajectory of the landslide particles, and the sliding process and accumulation area of the particles can be visualized in ArcGIS.

[0113] 3.2) Import the BIM model. ArcGIS supports direct import of BIM models in OBJ or FBX formats. Use the Add Data function to import the BIM model and ensure that the model has the correct georeferencing information (such as elevation and location) so that it can be aligned with the PFC model.

[0114] 3.3) Ensure that the BIM model and the landslide data imported from PFC share the same coordinate system (e.g., UTM coordinate system) so that the models can be correctly displayed and overlaid within the same spatial range. Use the 3D Analyst tool in ArcGIS to overlay the PFC model and BIM model. Ensure that both are displayed in the same 3D coordinate system and check the relative position of the landslide area to the highway and surrounding facilities. If particles from the PFC model slide onto areas such as roads and bridges in the BIM model, further analysis can be conducted on the potential impact of the landslide on infrastructure.

[0115] S8. Compare and analyze PFC and BIM models in ArcGIS to provide early warning of geological disasters

[0116] (1) Spatial analysis and visualization: Using ArcGIS's Hillshade, Slope and other analysis tools, we can analyze the relative relationship between the sliding path of landslide particles, the accumulation area and the facilities in the BIM model, whether the landslide will endanger the BIM model, or how large the landslide will affect the BIM. The Volume tool can be used to calculate the accumulation volume of the landslide and evaluate its impact on roads and bridges. In addition, different colors or transparencies can be set for the PFC landslide model to help clearly distinguish the landslide body from the roads, bridges and other structures in the BIM model. In the 3D Viewer, you can rotate and zoom the model to see how the landslide body contacts or accumulates with the infrastructure.

[0117] (3) Draw a schematic diagram of the analysis results. Use ArcGIS mapping tools to superimpose the PFC landslide model and the BIM road model to generate a two-dimensional or three-dimensional schematic diagram. Mark the sliding path of the landslide body, the intersection of the accumulation area and the road and bridge in the diagram, and analyze the threat level of the landslide to the highway facilities. Set different layer symbols and colors to distinguish the landslide body, road and other facilities, and use arrows to mark the movement direction of the landslide body. Finally, output the map. The final schematic diagram can be exported to PDF, JPEG or PNG format for reporting or displaying the risk analysis of landslides and roads. The generated three-dimensional schematic diagram can show how the landslide particles enter the highway area and provide visual risk warning information.

[0118] (4) Key points of the diagram:

[0119] PFC landslide position and movement path: The initial position, sliding path and final accumulation area of the landslide particles are shown in a schematic diagram.

[0120] BIM highway model: This model shows the designed structure of the highway, especially areas susceptible to landslides, such as roadbeds, bridges, and tunnel entrances.

[0121] Overlay Contrast: Clearly shows where landslide particles entered the highway, highlighting the infrastructure that may be affected by the accumulation of landslide material.

[0122] Color and symbols: Use different colors, transparency, and symbols to distinguish the PFC landslide model from the BIM facility model to ensure that the schematic diagram is clear and easy to understand.

[0123] Example 2

[0124] This study investigated slope stability analysis and risk warning for a 3-kilometer-long section of a highway in Hebei Province. By combining drone mapping, BIM modeling, and numerical simulation (ABAQUS and PFC), they achieved a landslide risk warning.

[0125] Step S1: UAV mapping and data collection

[0126] First, a 3-kilometer section of a highway in Hebei Province with potential landslide risk was selected, focusing on the roadbed, bridges, and nearby hillside areas. Low-altitude mapping was conducted using a drone equipped with LiDAR and a high-resolution camera.

[0127] Flight parameter settings: The drone was set to fly at an altitude of 120 meters to cover the terrain of the 3-kilometer section, with lateral overlap set to 70% and forward overlap set to 80% to ensure mapping accuracy. The flight speed was set to 5 meters per second to obtain high-precision imagery and raw point cloud data.

[0128] Data Collection: The main data collected by the drone during flight include raw point cloud data (stored in LAS format), high-resolution image data (orthophotos), and digital elevation models (DEM, stored in GeoTIFF format).

[0129] Step S2: Data processing and 3D modeling

[0130] Use Context Capture software to process the raw point cloud data collected by the drone: Import the LAS format point cloud data obtained by the drone into Context Capture and perform data cleaning to remove noise points, duplicate points, and erroneous data.

[0131] Step S3: Based on the processed point cloud data, Smart 3D is used to generate a detailed, realistic 3D terrain, including features such as hillsides, bridges, and roadbeds along the road section. This terrain is then exported to STL and OBJ formats for subsequent numerical simulation and modeling.

[0132] Step S4: BIM model establishment

[0133] Import the point cloud data pre-processed in step 2 into the Revit software for 3D BIM modeling operations. Use Autodesk Revit software to create a 3D BIM model of the road section. Use the terrain modeling tool in Revit to generate and build infrastructure models of roads, bridges, and drainage systems based on the highway design data. Focus on modeling key structures such as the roadbed, bridges, and culverts of the highway to ensure that the geometric accuracy of the model is consistent with the actual situation. At the same time, combine the terrain features in the point cloud data (such as trees, valleys, etc.) to refine the model details to avoid inaccurate modeling due to occlusion or errors. Through comparative analysis of the terrain features in the point cloud data and the infrastructure model, verify and optimize the size and position of the BIM model to ensure that it fully matches the actual terrain and facility environment, and then obtain a BIM model of the in-service road and surrounding houses based on the point cloud data;

[0134] Step S5: Compare and adjust the BIM model with the high-precision real-life 3D terrain generated by Smart3D software to ensure seamless integration of the facilities in the BIM model with the terrain in the real-life 3D terrain. Then, integrate the BIM model with the high-precision real-life 3D terrain generated by Smart3D software and optimize the geometric details of the BIM model to ensure that the BIM meets design requirements and truly reflects actual terrain conditions. Export the BIM model that meets the requirements in Revit to a format suitable for processing in ArcGIS software.

[0135] At this point, the BIM model simultaneously integrates the road design data and on-site terrain characteristics, providing a basis for subsequent analysis.

[0136] Step S6: Numerical simulation analysis

[0137] Export the BIM model processed in step 5 from Revit in STL format for ABAQUS numerical simulation and analysis, and export it from Revit in DXF format for PFC numerical simulation and analysis;

[0138] (1) In order to analyze the stability of the slopes around the road section under different rainfall conditions, finite element analysis was performed using ABAQUS:

[0139] First, the terrain portion of the BIM model (STL format) was imported into ABAQUS, and a 2D cross-section model was generated to simulate the key geometric features of the slope. The imported slope geometry was then meshed using four-node plane strain elements (CPE4P elements). The mesh was refined in key areas (such as the interface between the landslide and the roadbed) to improve simulation accuracy.

[0140] Material parameter setting: According to geological exploration data, the physical parameters of the hillside soil in this section are: Soil density: 1600kg / m 3; Internal friction angle: 30°; Cohesion: 20kPa; Deformation modulus: 5×107Pa; Poisson's ratio: 0.3;

[0141] Rainfall condition simulation: set a 50-year rainfall condition with a rainfall of 120 mm and a duration of 3 days, set the pore water pressure boundary condition, and simulate the impact of rainfall infiltration on slope stress and displacement.

[0142] Analysis results: The simulation results show that when the rainfall continues for 48 hours, the stress on the outer edge of the slope concentrates, with the maximum stress being 120kPa. The maximum displacement is located below the roadbed, reaching 0.6 meters. The plastic zone develops to the middle of the slope, and the landslide risk increases.

[0143] (2) In order to further simulate the movement process of landslide particles, discrete element analysis was performed using PFC software:

[0144] The DXF format exported from Revit was imported into PFC2D to generate a two-dimensional cross-sectional model of the landslide. Discrete element analysis was performed on the potentially unstable areas using PFC software to dynamically simulate the movement of particles during the landslide. Dynamic conditions (such as rainfall and earthquakes) were applied to simulate the sliding path, accumulation range, and displacement of particles, as well as their kinematic characteristics under earthquake action.

[0145] Specifically, the porosity of the sliding zone was set to 0.25 for particle generation, and a random generation method was used to generate landslide particles with a particle radius of 0.5-0.7 meters. The same method was used to generate smaller particles in the sliding bed with a particle radius of 0.3-0.5 meters. A total of about 8,000 particles were generated to simulate the landslide process. For the rainfall load conditions, the same rainfall load as ABAQUS was applied, and the pore water pressure boundary was set to simulate the effect of rainfall on particle movement. The final PFC simulation results showed that 48 hours after the rainfall, the particles on the upper part of the sliding body began to slide unstably, with a maximum displacement of 1.2 meters. The landslide particles eventually accumulated at the edge of the highway, and some particles invaded the roadbed.

[0146] S7: Comprehensive analysis and risk warning

[0147] ArcGIS software was used for visualization analysis and risk warning. The simulation results from ABAQUS and PFC were exported to STL, DXF, and CSV formats, respectively, and then imported into ArcGIS. The stress and displacement fields generated by ABAQUS were overlaid with the PFC landslide particle motion paths for analysis. Using ArcGIS's 3D analysis tools, the BIM model, the slope stress field, and the landslide particle model were then overlaid to analyze the impact of the landslide on the highway.

[0148] Analysis shows that in the most severe cases, some of the accumulated debris could affect the highway subgrade. Combining the analysis results from ABAQUS and PFC, safety thresholds for stress and displacement were set. When the slope displacement reaches 0.5 meters or the pore water pressure exceeds 30 kPa, the system triggers an alert, alerting maintenance departments to initiate emergency response. Furthermore, based on the simulated impact range of the debris flow, users can accurately determine which sections of highway are at high and low risk. Similarly, the risk level of homes along the highway can be determined and differentiated.

[0149] However, in Figure 3 Without this patented technology, users can only roughly determine that the terrain along the highway is unstable and that disasters such as landslides and mudslides are likely to occur. However, they cannot accurately predict which sections of the road are at high risk of geological disasters and which are at low risk. Consequently, relevant management departments are unable to issue accurate early warning information. For safety reasons, all sections of the road are typically designated as high-risk, which consumes a lot of manpower and resources.

[0150] The present invention obtains high-precision original point cloud data and image data of the target highway and surrounding areas through close photogrammetry by unmanned aerial vehicle. The data after high-precision data processing is imported into BIM software (such as Revit, Navisworks) to generate BIM models of roads, bridges and surrounding mountains to ensure the terrain compatibility and structural integrity of the model. The BIM model is exported from Revit in OBJ format and imported into ABAQUS for finite element analysis, and the physical properties and boundary conditions of the slope are set to simulate the stress distribution and displacement response under rainfall conditions. At the same time, PFC (particle flow simulation tool) is used to generate a particle model of the landslide, and the sliding path, accumulation range and displacement of the particles are simulated by applying dynamic conditions (such as rainfall, earthquake, etc.). After the analysis is completed, the numerical simulation results are imported into ArcGIS, and three-dimensional overlay analysis is performed with the BIM model to generate a visual risk warning map, which provides a scientific basis for maintenance and decision-making. There are differences between the terrain data, BIM model, ABAQUS stress field, PFC landslide particle flow and other multidimensional data collected by the drone. This application superimposes these data on the ArcGIS platform and realizes unified display through three-dimensional visualization analysis. This integration method intuitively displays the spatial relationship between landslide areas and road facilities, enabling risk warning to have multi-dimensional information integration capabilities.

[0151] Any matters not described in the present invention are applicable to the prior art.

Claims

1. A highway geological disaster risk early warning method based on BIM and numerical simulation, characterized by: The steps of the early warning method are: Step 1: Use drone close-up photogrammetry to obtain high-precision original point cloud data and image data along the target highway and surrounding areas; Step 2: Import the raw point cloud data collected by the drone into Context Capture software. Context Capture then performs denoising, optimization, and precision adjustment on the raw point cloud data, removing invalid and duplicate points. Data from different flight batches is then stitched together to generate a complete 3D point cloud model. The spatial coordinate system is then adjusted based on project requirements to ensure consistency with the coordinates of the subsequently established BIM model, completing the preprocessing of the point cloud data. Step 3: Based on the data pre-processed in step 2, Smart 3D software is used to generate high-precision real-life 3D terrain of the in-service road and surrounding areas, and the real-life 3D terrain data is converted into a data format; Step 4: Import the point cloud data pre-processed in Step 2 into Revit software for 3D BIM modeling. Use Revit's terrain modeling tools to generate a terrain surface that accurately reflects the actual terrain features of the highway and its surrounding areas. Based on the existing design data of the highway infrastructure, use Revit to create a highway infrastructure model. At the same time, combine the terrain features in the point cloud data to refine the model details and avoid inaccurate modeling due to occlusion or errors. Through comparative analysis of the terrain features in the point cloud data and the infrastructure model, verify and optimize the size and position of the BIM model to ensure that it fully matches the actual terrain and facility environment, thereby obtaining a BIM model of the in-service road and surrounding houses based on the point cloud data. Step 5: Compare and adjust the BIM model with the high-precision real-world 3D terrain generated by Smart3D software to ensure seamless integration of the facilities in the BIM model and the terrain in the real-world 3D terrain. Then, integrate the BIM model with the high-precision real-world 3D terrain generated by Smart3D software and optimize the geometric details of the BIM model to ensure that the BIM model meets design requirements and truly reflects actual terrain conditions. In Revit, export the BIM model that meets the requirements into a format suitable for processing in ArcGIS software. Step 6: Export the BIM model processed in step 5 from Revit to STL format or OBJ format for ABAQUS numerical simulation and analysis. The STL format data is used for three-dimensional numerical simulation modeling and analysis, and the OBJ format is used for two-dimensional section numerical simulation modeling and analysis. Export the DXF format from Revit to PFC numerical simulation and analysis. A complete slope geometry model was established using ABAQUS software for finite element analysis. The physical properties and boundary conditions of the slope were set to simulate the slope stress and displacement fields under rainfall conditions. When the convergence conditions were reached, the area of potential instability, i.e., the instability range, was identified. The DXF file exported from Revit was imported into PFC to generate a landslide model. The sliding mass and sliding bed of the slope were divided according to the instability range under rainfall conditions determined using ABAQUS. Discrete element analysis was performed on the potentially unstable areas using PFC software to dynamically simulate the movement of particles during the landslide. Dynamic conditions were applied to simulate the sliding path, accumulation range, and displacement of particles, as well as their kinematic characteristics under earthquake action. Step 7: Export the numerical simulation results of PFC into a format suitable for processing by ArcGIS software and import it into the ArcGIS platform. Simultaneously, import the BIM model from step 5 into the ArcGIS platform. Set the BIM model and the landslide data imported from PFC to share the same coordinate system so that they can be correctly displayed and three-dimensionally superimposed within the same spatial range. Step 8: Use ArcGIS analysis tools to analyze the relative relationship between the sliding path of landslide particles, the accumulation area, and the facilities in the BIM model, and intuitively display the interaction between the landslide area, accumulation location, and road facilities. Ultimately, form early warning analysis and maintenance recommendations, and generate a visual risk warning map.

2. The method according to claim 1, characterized in that The method can provide dynamic early warning. The data collected by the drone communicates with the Context Capture software. The Context Capture software communicates with the Smart 3D software and the Revit software respectively. The Revit software communicates with the ABAQUS software and the PFC software respectively. At the same time, the ABAQUS software communicates with the PFC software, and the Revit software and the PFC software both communicate with the ArcGIS platform. The drone dynamically collects data, so that the BIM model can be continuously updated according to the periodic data collected by the drone, and early warning adjustments are made in combination with the simulation results of the PFC, thereby maintaining the dynamicity and accuracy of the BIM model and ensuring that the early warning information always conforms to the current environment.

3. The method according to claim 2, characterized in that The ABAQUS software also communicates with the ArcGIS platform to display the numerical simulation results of the ABAQUS software.

4. The method according to claim 1, wherein In step 5, the facilities in the BIM model include roads, bridges, tunnels and surrounding mountains.

5. The method according to claim 1, wherein The process of processing point cloud data with the help of Smart 3D software includes: matching points with the same name using a feature-based matching algorithm, performing multi-view image joint adjustment using the bundle method regional block adjustment algorithm, and completing the preliminary construction of a high-density point cloud based on the clustering algorithm and the patch-based dense matching algorithm; finally, the TIN network model constructed based on the point cloud is combined with texture mapping to achieve the construction of a high-precision real-life 3D model.

6. A highway geological disaster risk early warning system based on BIM and numerical simulation, characterized by: The method of claim 1 is performed, wherein the system comprises: The UAV data acquisition module uses drones to collect high-precision raw point cloud data and image data along the target highway and surrounding areas; A data preprocessing module is used to preprocess the data collected by the UAV data acquisition module; The BIM modeling module is used to perform BIM modeling based on pre-processed data and optimize the BIM model using the real-life 3D terrain generated by Smart 3D software, thereby obtaining a BIM model that meets design requirements and truly reflects actual terrain conditions. The numerical simulation module is used to perform finite element analysis and discrete element analysis based on the BIM model output by the BIM modeling module. The finite element analysis is used to determine the area of possible instability when the convergence conditions are reached. The discrete element analysis is used to determine the sliding path, accumulation range and displacement of particles in the area of possible instability, as well as the kinematic characteristics under earthquake action. The GIS data integration module is used to integrate and display the output results of the BIM modeling module and the output results of the numerical simulation module, analyze the relative relationship between the sliding path of landslide particles, the accumulation area and the facilities in the BIM model, and intuitively display the interaction between the landslide area, accumulation location and road facilities; The risk warning and feedback module is used to conduct risk assessment based on the output of the GIS data integration module, and generate a visual risk warning map based on the final warning analysis and maintenance recommendations.

Images