Bridge model establishment method based on multi-source point cloud data

Abstract

The invention discloses a bridge model establishment method based on multi-source point cloud data. The method comprises the following steps: S1, establishing a bridge BIM model according to a bridge design drawing; s2, collecting geographical environment information of a construction site, and establishing a three-dimensional environment model; s3, importing the BIM model and the three-dimensional environment model into a digital twinborn platform, integrating the BIM model and the three-dimensional environment model based on the digital twinborn platform, and simulating bridge construction by using finite element simulation software based on the BIM model and the three-dimensional environment model to obtain a bridge construction simulation model; s4, acquiring bridge construction image data in real time from different visual angles, respectively establishing bridge point cloud models, and registering a plurality of bridge point cloud models to obtain a three-dimensional point cloud model; and S5, importing the three-dimensional point cloud model into a digital twin platform, and integrating the bridge construction simulation model and the three-dimensional point cloud model by using the digital twin platform to obtain a bridge dynamic construction model. According to the bridge model established by the invention, comprehensive perception, fidelity analysis and real-time monitoring can be carried out on the construction condition, and the construction quality and the construction safety are ensured.

Description

Technical Field

[0001] The present invention relates to the field of bridge construction, and more specifically, to a method for establishing a bridge model based on multi-source point cloud data. Background Art

[0002] As more and more bridge engineering projects have put forward higher requirements for construction monitoring. At present, bridge construction monitoring generally uses BIM technology to build a virtual model of the structure to manage various types of information, but it mainly focuses on information management rather than information application. The virtual model built based on BIM technology cannot obtain the on-site construction situation and changing trends in real time, cannot fully perceive, analyze and control the construction with fidelity, and cannot guarantee the construction quality and safety.

[0003] To solve the above problems, it is necessary to provide a new bridge model establishment method based on multi-source point cloud data to comprehensively perceive, analyze and monitor the construction situation in real time to ensure construction quality and safety. Summary of the invention

[0004] Another object of the present invention is to provide a method for establishing a bridge model based on multi-source point cloud data. The bridge model established by this method can perform comprehensive perception, fidelity analysis and real-time monitoring of the construction situation, thereby ensuring construction quality and construction safety.

[0005] In order to achieve these purposes and other advantages according to the present invention, a bridge model establishment method based on multi-source point cloud data is provided, comprising: S1. Establish a bridge BIM model according to the bridge design drawings; S2. Collect geographical environment information of the construction site and establish a three-dimensional environment model; S3, importing the BIM model and the three-dimensional environment model into the digital twin platform, integrating the BIM model and the three-dimensional environment model based on the digital twin platform, and simulating the bridge construction using finite element simulation software based on this to obtain a bridge construction simulation model; S4, acquiring bridge construction image data from different perspectives in real time, and establishing bridge point cloud models respectively, and registering multiple bridge point cloud models to obtain a three-dimensional point cloud model; S5. Import the 3D point cloud model into the digital twin platform, and use the digital twin platform to integrate the bridge construction simulation model and the 3D point cloud model to obtain a dynamic construction model of the bridge.

[0006] Preferably, step S3 further includes: A. Count the construction structure information of the same type of bridges in historical bridge construction projects to obtain historical construction structure information data; B. Installing a measuring device on the bridge component, obtaining the bridge construction structure information in real time through the measuring device, and obtaining real-time construction structure information data; C. Collect historical construction structure information data and real-time construction structure information data as training sets, train neural network model parameters, and obtain a construction structure information prediction model; In step S5, the construction structure information prediction model and the three-dimensional point cloud model are imported into the digital twin platform, and the digital twin platform is used to integrate the bridge construction simulation model, the construction structure information prediction model and the three-dimensional point cloud model to obtain a dynamic construction model of the bridge.

[0007] Preferably, the construction structure information includes structural stress information, structural internal force information, temperature characteristic information and geographical location information.

[0008] Preferably, step S4 specifically includes: S41, extracting the origin of the BIM model, placing a calibration target at the construction site corresponding to the origin of the BIM model based on the physical space information of the bridge construction, the geographical location information, and the physical space information of the structures around the construction site, using a 3D laser scanner with the calibration target as the scanning reference to obtain multiple image data of the bridge construction from different perspectives, and respectively establishing bridge point cloud models; S42, preprocessing the multiple bridge point cloud models, and registering the multiple preprocessed bridge point cloud models to obtain a three-dimensional point cloud model.

[0009] Preferably, step S42 specifically includes: S421, filtering and denoising multiple bridge point cloud models; S422, based on the topological relationship of the bridge, prior engineering information, and multiple construction image data, the point clouds of the multiple bridge point cloud models after filtering and denoising are divided into multiple groups of common point clouds and multiple groups of discrete point clouds based on a machine learning algorithm, wherein the common point cloud is the point cloud of the overlapping area of ​​the multiple bridge point cloud models; S423. Based on a point cloud registration algorithm, multiple groups of public point clouds are registered to obtain a basic three-dimensional point cloud model, and the multiple groups of discrete point clouds are fused with the basic three-dimensional point cloud model to obtain a three-dimensional point cloud model.

[0010] Preferably, the fusion of multiple sets of discrete point clouds with the basic three-dimensional point cloud model specifically includes the following steps: a. Extract the point cloud located at the edge of the basic 3D point cloud model as a training sample and input it into the full-precision network to train the full-precision network to obtain a trained full-precision network; b. Use the trained full-precision network to train multiple groups of discrete point clouds to obtain trained discrete point clouds; c. Perform axis fitting on the basic point cloud model, project the trained discrete point cloud onto the plane according to the axis direction, and use the least squares method to perform surface fitting to obtain the fitting surface; b. Fuse the fitted surface with the basic 3D point cloud model to obtain a 3D point cloud model.

[0011] The present invention includes at least the following beneficial effects: the present invention utilizes the digital twin platform to integrate the bridge construction simulation model and the three-dimensional point cloud model to obtain a dynamic bridge construction model. Through the dynamic bridge construction model, based on the digital twin technology, the synchronization of the physical world and the virtual world can be achieved, and the simulation and analysis capabilities of the construction process can be enhanced. During the construction process, by selecting the corresponding bridge components in the dynamic bridge construction model, the bridge construction data can be viewed in real time to obtain the on-site construction data and changing trends in real time, and to achieve comprehensive perception, fidelity analysis and real-time control of construction quality and safety information, and to provide precise control for information monitoring and construction decision-making of bridge construction in all aspects, thereby improving construction accuracy and timeliness, thereby facilitating managers to judge the construction quality and safety, perceive and prevent problems that may arise during construction in advance, and facilitate on-site managers to make adaptive adjustments to the construction plan in a timely manner to ensure construction quality and construction safety.

[0012] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. DETAILED DESCRIPTION

[0013] The present invention is described in further detail below so that those skilled in the art can implement it according to the description.

[0014] It should be noted that, in the description of the present invention, terms such as "lateral", "longitudinal", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" to indicate directions or positional relationships are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0015] The present invention provides a bridge model establishment method based on multi-source point cloud data, which is applied to the newly built Hefei-Wuhan high-speed railway Hubei section HWZQ-5 section on the main line across the Han-Shi Expressway and ramp (52+92+80+86+52)m rigid frame continuous beam engineering project, specifically including the following steps: S1. Establish a bridge BIM model according to the bridge design drawings; S2. Collect geographical environment information of the construction site and establish a three-dimensional environment model; Specifically, the geographical environment data of the construction site is collected by drones, and after being summarized, analyzed and compared with the design drawings by project managers, a three-dimensional environmental model is established; S3, importing the BIM model and the three-dimensional environment model into the digital twin platform, integrating the BIM model and the three-dimensional environment model based on the digital twin platform, and simulating the bridge construction using finite element simulation software based on this to obtain a bridge construction simulation model; The construction simulation model incorporates a three-dimensional environment model, making the construction simulation closer to the actual situation on the construction site, and providing a basis for the accuracy and authenticity of the bridge construction dynamic construction model's perception, analysis, and real-time monitoring of the construction situation; S4, acquiring bridge construction image data from different perspectives in real time, and establishing bridge point cloud models respectively, and registering multiple bridge point cloud models to obtain a three-dimensional point cloud model; S5. Import the three-dimensional point cloud model into the digital twin platform, and use the digital twin platform to integrate the bridge construction simulation model and the three-dimensional point cloud model to obtain a dynamic construction model of the bridge.

[0016] In the above technical scheme, the present invention uses a digital twin platform to integrate a bridge construction simulation model and a three-dimensional point cloud model to obtain a dynamic construction model of the bridge. Through the dynamic construction model of the bridge, based on the digital twin technology, the synchronization of the physical world and the virtual world can be achieved, and the simulation and analysis capabilities of the construction process can be enhanced. During the construction process, by selecting corresponding bridge components in the dynamic construction model of the bridge, the bridge construction data can be viewed in real time to obtain real-time on-site construction data and changing trends, and to achieve comprehensive perception, fidelity analysis and real-time control of construction quality and safety information, and provide precise control for information monitoring and construction decision-making of bridge construction in all aspects, improve construction accuracy and timeliness, thereby facilitating managers to judge the construction quality and safety, perceive and prevent possible problems in construction in advance, and facilitate on-site managers to make adaptive adjustments to the construction plan in a timely manner to ensure construction quality and construction safety.

[0017] In another technical solution, after step S3, the following steps are further included: A. Count the construction structure information of the same type of bridges in historical bridge construction projects to obtain historical construction structure information data; B. Install a measuring device on the bridge component, obtain the bridge construction structure information in real time through the measuring device, and obtain real-time construction structure information data; the construction structure information includes structural stress information, structural internal force information, temperature characteristic information and geographical location information, and the measuring device is a sensor; C. Collect historical construction structure information data and real-time construction structure information data as training sets, train neural network model parameters, and obtain a construction structure information prediction model; In step S5, the construction structure information prediction model and the three-dimensional point cloud model are imported into the digital twin platform, and the digital twin platform is used to integrate the bridge construction simulation model, the construction structure information prediction model and the three-dimensional point cloud model to obtain a dynamic construction model of the bridge.

[0018] In this technical solution, a construction structure information prediction model is established and integrated with the bridge construction simulation model and the three-dimensional point cloud model. The factors affecting construction quality and safety, such as construction stress, internal force, and temperature, are taken into account. The dynamic construction model of the bridge is used to view the construction structure information, position status and other data of the bridge in real time. The structural stress, structural internal force, temperature data, and geographic location information are linked to the construction monitoring operation in real time, which further improves the comprehensiveness of the monitoring data and greatly improves the accuracy of the monitoring data.

[0019] In another technical solution, step S4 specifically includes: S41, extracting the origin of the BIM model, placing a calibration target at the construction site corresponding to the origin of the BIM model based on the physical space information of the bridge construction, the geographical location information, and the physical space information of the structures around the construction site, using a 3D laser scanner with the calibration target as the scanning reference to obtain multiple image data of the bridge construction from different perspectives, and respectively establishing bridge point cloud models; S42, preprocessing the multiple bridge point cloud models, and registering the multiple preprocessed bridge point cloud models to obtain a three-dimensional point cloud model.

[0020] In this technical solution, by acquiring multiple point cloud models of the bridge from different perspectives, the comprehensiveness of the acquired point cloud data can be guaranteed, and then the multiple bridge point cloud models are aligned to ensure the integrity of the established three-dimensional point cloud model, so as to ensure the accuracy of subsequent monitoring data; the calibration target is arranged at the position corresponding to the origin of the BIM model at the construction site, which can further ensure the consistency of the three-dimensional point cloud model and the BIM model, and facilitate the integration of the construction simulation model and the three-dimensional point cloud model.

[0021] In another technical solution, step S42 specifically includes: S421, filtering and denoising multiple bridge point cloud models; S422, based on the topological relationship of the bridge, prior engineering information, and multiple construction image data, the point clouds of the multiple bridge point cloud models after filtering and denoising are divided into multiple groups of common point clouds and multiple groups of discrete point clouds based on a machine learning algorithm, wherein the common point cloud is the point cloud of the overlapping area of ​​the multiple bridge point cloud models; S423, based on the point cloud registration algorithm, registering the multiple groups of public point clouds to obtain a basic three-dimensional point cloud model, and fusing the multiple groups of discrete point clouds with the basic three-dimensional point cloud model to obtain a three-dimensional point cloud model; The fusion of multiple discrete point clouds and the basic 3D point cloud model specifically includes the following steps: a. Extract the point cloud located at the edge of the basic 3D point cloud model as a training sample and input it into the full-precision network to train the full-precision network to obtain a trained full-precision network; b. Use the trained full-precision network to train multiple groups of discrete point clouds to obtain trained discrete point clouds; c. Perform axis fitting on the basic point cloud model, project the trained discrete point cloud onto the plane according to the axis direction, and use the least squares method to perform surface fitting to obtain the fitting surface; b. Fuse the fitted surface with the basic 3D point cloud model to obtain a 3D point cloud model.

[0022] In this technical solution, when using a 3D laser scanner to obtain bridge construction image data, the edge part will be incomplete, resulting in the missing edge point cloud in the established bridge point cloud model. Therefore, in order to improve the accuracy of point cloud registration and the completeness of the 3D point cloud model after registration, the point cloud in each bridge point cloud model is divided into a public point cloud and a discrete point cloud. The public point cloud is first registered to ensure the accuracy of the point cloud in the overlapping area. The discrete point cloud is then trained through a full-precision network to achieve the extension of the edge (discrete) point cloud from low precision to high precision. The high-precision discrete point cloud is then projected based on the axis of the basic point cloud model. , thereby establishing a mapping relationship between the discrete point cloud and the basic point cloud model, based on the mapping relationship, realizing the fusion of the surface fitted by the high-precision discrete point cloud and the basic three-dimensional point cloud model, thereby repairing the missing parts of the basic three-dimensional point cloud model, and finally obtaining a high-precision and complete three-dimensional point cloud model; the bridge edge line refers to the edge line of the bridge, which is also used to control the line shape of the bridge to a certain extent. Therefore, the bridge edge line is very important in bridge design and construction. This application adopts the point cloud division method to align and fuse the point clouds in batches, ensuring the integrity and accuracy of the point cloud located on the edge line of the bridge, and ensuring the availability and reliability of the three-dimensional point cloud model.

[0023] Although the embodiments of the present invention have been disclosed as above, they are not limited to the applications listed in the specification and the implementation modes. They can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily implemented. Therefore, without departing from the general concept defined by the claims and the scope of equivalents, the present invention is not limited to specific details.

Claims

1. A bridge modeling method based on multi-source point cloud data, characterized in that: include: S1. Establish a bridge BIM model according to the bridge design drawings; S2. Collect geographical environment information of the construction site and establish a three-dimensional environment model; S3, importing the BIM model and the three-dimensional environment model into the digital twin platform, integrating the BIM model and the three-dimensional environment model based on the digital twin platform, and simulating the bridge construction using finite element simulation software based on this to obtain a bridge construction simulation model; S4, acquiring bridge construction image data from different perspectives in real time, and establishing bridge point cloud models respectively, and registering multiple bridge point cloud models to obtain a three-dimensional point cloud model; S5. Import the 3D point cloud model into the digital twin platform, and use the digital twin platform to integrate the bridge construction simulation model and the 3D point cloud model to obtain a dynamic construction model of the bridge.

2. The bridge modeling method based on multi-source point cloud data according to claim 1, characterized in that: After step S3, the following steps are also included: A. Count the construction structure information of the same type of bridges in historical bridge construction projects to obtain historical construction structure information data; B. Installing a measuring device on the bridge component, obtaining the bridge construction structure information in real time through the measuring device, and obtaining real-time construction structure information data; C. Collect historical construction structure information data and real-time construction structure information data as training sets, train neural network model parameters, and obtain a construction structure information prediction model; In step S5, the construction structure information prediction model and the three-dimensional point cloud model are imported into the digital twin platform, and the digital twin platform is used to integrate the bridge construction simulation model, the construction structure information prediction model and the three-dimensional point cloud model to obtain a dynamic construction model of the bridge.

3. The bridge modeling method based on multi-source point cloud data according to claim 2, characterized in that: The construction structure information includes structural stress information, structural internal force information, temperature characteristic information and geographical location information.

4. The bridge modeling method based on multi-source point cloud data according to claim 3, characterized in that: Step S4 specifically includes: S41, extracting the origin of the BIM model, placing a calibration target at the construction site corresponding to the origin of the BIM model based on the physical space information of the bridge construction, the geographical location information, and the physical space information of the structures around the construction site, using a 3D laser scanner with the calibration target as the scanning reference to obtain multiple image data of the bridge construction from different perspectives, and respectively establishing bridge point cloud models; S42, preprocessing the multiple bridge point cloud models, and registering the multiple preprocessed bridge point cloud models to obtain a three-dimensional point cloud model.

5. The bridge model building method based on multi-source point cloud data according to claim 4, characterized in that: Step S42 specifically includes: S421, filtering and denoising multiple bridge point cloud models; S422, based on the topological relationship of the bridge, prior engineering information, and multiple construction image data, the point clouds of the multiple bridge point cloud models after filtering and denoising are divided into multiple groups of common point clouds and multiple groups of discrete point clouds based on a machine learning algorithm, wherein the common point cloud is the point cloud of the overlapping area of ​​the multiple bridge point cloud models; S423. Based on a point cloud registration algorithm, multiple groups of public point clouds are registered to obtain a basic three-dimensional point cloud model, and the multiple groups of discrete point clouds are fused with the basic three-dimensional point cloud model to obtain a three-dimensional point cloud model.

6. The bridge model building method based on multi-source point cloud data according to claim 5, characterized in that: The fusion of multiple discrete point clouds and the basic 3D point cloud model specifically includes the following steps: a. Extract the point cloud located at the edge of the basic 3D point cloud model as a training sample and input it into the full-precision network to train the full-precision network to obtain a trained full-precision network; b. Use the trained full-precision network to train multiple groups of discrete point clouds to obtain trained discrete point clouds; c. Perform axis fitting on the basic point cloud model, project the trained discrete point cloud onto the plane according to the axis direction, and use the least squares method to perform surface fitting to obtain the fitting surface; b. Fuse the fitted surface with the basic 3D point cloud model to obtain a 3D point cloud model.