Electromechanical pipeline installation condition monitoring system based on BIM and AR

Abstract

The invention discloses an electromechanical pipeline installation condition monitoring system based on BIM and AR, and relates to the technical field of pipeline installation. The electromechanical pipeline installation condition monitoring system based on BIM and AR comprises a three-dimensional scanning and data preprocessing module, a drawing three-dimensional modeling module, an AR distinguishing display and error identification module, a correction suggestion and real-time guidance module and a recording and feedback improvement module. Through three-dimensional scanning and data preprocessing, high-precision actual pipeline data acquisition is realized, and a solid foundation is provided for subsequent analysis. And by introducing the drawing three-dimensional modeling module, the comparison between the design and the actual installation is more intuitive. The application of the AR distinguishing display and error recognition module not only improves the accuracy of error recognition, but also predicts an error threshold value through a regression model, and effectively evaluates the severity of errors. And the correction suggestion and real-time guidance module can provide correction suggestions in real time to guide constructors to operate, so that the construction efficiency and quality are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of pipeline installation, and in particular to a BIM and AR-based electromechanical pipeline installation monitoring system. Background Art

[0002] The electromechanical systems of high-rise or super high-rise buildings usually include water supply, hot water, drainage, fire protection, air conditioning, heating, smoke exhaust, lighting, power, lightning protection, and intelligence systems. The number of equipment and pipelines required for these systems is huge, and they are intertwined and related like a spider web, making the comprehensive layout and construction of pipelines difficult.

[0003] At present, for the construction of electromechanical pipelines, strict construction is required in accordance with the drawings. However, in the actual construction process, errors are still prone to occur, and due to the influence of actual construction conditions, the construction cannot be carried out completely according to the drawings. Under the accumulation of errors, large construction errors will eventually be caused; Publication No. CN116050040A discloses an intelligent layout method and system based on the spatial characteristics of pipeline layout, the method includes the following steps: S1, measuring the architectural parameters of the construction building to be arranged for the pipeline and constructing a digital twin scene; S2, importing the original pipeline layout drawings into the digital twin scene to construct a pipeline layout model; S3, obtaining the process parameters of the pipeline layout to evaluate and optimize the pipeline layout model; S4, obtaining the pipeline layout spatial characteristics and construction sequence of the optimized pipeline layout model; S5, performing pipeline layout construction according to the pipeline layout spatial characteristics and construction sequence; S6, measuring the actual spatial characteristics of the pipeline during the construction process to monitor the construction process errors; S7, dynamically adjusting and correcting the pipeline layout construction process according to the monitoring results.

[0004] As shown in the above technology, there are now more professional systems to replace drawings for viewing in order to carry out more accurate pipeline construction. However, how to intuitively view the difference between the actual construction and the drawings, and how to intuitively distinguish the error value for correction still needs to be improved. Failure to promptly distinguish and correct the error will affect the subsequent construction accuracy. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a BIM and AR-based electromechanical pipeline installation monitoring system, which solves the problem that the prior art cannot intuitively view the difference between actual construction and drawings, and cannot intuitively distinguish error values ​​for correction.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: A BIM and AR-based electromechanical pipeline installation monitoring system, comprising:

[0007] Drawing 3D modeling module, used to create 3D drawing models of electromechanical pipelines based on design drawings;

[0008] The actual pipeline 3D modeling module uses the AR device shooting function to capture images of the actual pipeline during the actual installation process, processes the images to extract point cloud data of the actual pipeline, generates 3D scanning data through the pre-processing steps of denoising, registration, and splicing, and creates a model of the actually installed pipeline based on the 3D scanning data;

[0009] AR difference display and error identification module, used to import 3D drawing modeling and actual modeling into AR devices, distinguish and display the two models, calculate the spatial distance to identify errors, optimize the error calculation method to improve accuracy, and establish a regression model based on historical data training to predict threshold comparison errors and judge the severity of errors; at the same time, it provides error marking and correction suggestion display functions;

[0010] The correction suggestion and real-time guidance module classifies and quantifies the identified errors, defines the size and importance of each error, generates correction suggestions and displays them in real time on the AR device to guide construction personnel to make adjustments;

[0011] The recording and feedback improvement module records the pipeline spatial feature data before and after correction and the error data identified during the correction process, and records the correction suggestion data as historical data to train and improve the regression model.

[0012] Preferably, the work content of the drawing three-dimensional modeling module specifically includes: according to the design drawings, using BIM software to create a three-dimensional drawing model of the electromechanical pipeline, wherein the parameters include defining the type, size, position and connection relationship of the pipeline.

[0013] Preferably, the working contents of the actual pipeline 3D modeling module specifically include:

[0014] Point cloud data extraction: Select an AR device with a depth sensor to obtain depth information through other structured light or stereo vision methods, extract depth data from the image and generate point cloud data;

[0015] Data preprocessing:

[0016] Denoising: removing unwanted noise from scan data;

[0017] Registration: Use the ICP algorithm to align images from multiple perspectives into a common coordinate system;

[0018] Stitching: Use the point cloud stitching algorithm to merge the registered multi-view point cloud data into a complete 3D model data;

[0019] Actual modeling: import the preprocessed 3D model data into the modeling software, generate polygonal meshes or surfaces from the point cloud data, and create pipeline models based on the meshes or surfaces.

[0020] Preferably, the specific steps of the actual modeling include:

[0021] A1. Point cloud data import and preparation:

[0022] Import point cloud data: Use modeling software to import pre-processed point cloud data, denoted as P = {p1, p2, ..., p n}, where p i Represents the three-dimensional coordinates of the i-th point;

[0023] Data preparation: Clean the point cloud data and remove duplicate points and isolated points, denoted as P c =C(P); and according to the project requirements, the point cloud data is segmented and recorded as P s =S(P c );

[0024] A2. Automatic or semi-automatic mesh generation:

[0025] Automatic mesh generation: Use the built-in meshing tool in the software to convert point cloud data into polygonal meshes, denoted as M=Mesh(P s ), set the grid parameters including resolution r and smoothness s;

[0026] Semi-automatic adjustment: Use editing tools to manually adjust the mesh shape of the automatically generated mesh, denoted as M a =Adjust(M);

[0027] A3. Pipeline information extraction and identification:

[0028] Pipeline recognition based on machine learning: The pipeline recognition algorithm based on machine learning is applied to process the grid data, denoted as I=Recognize(M a ), where I represents the extracted pipeline information;

[0029] When training the model, the labeled pipeline data is used as the training set, denoted as D = {(a1, b1), (a2, b2), ..., (a m , b m )}, where a k represents the input feature, b k Indicates the marked pipeline information;

[0030] Manual verification and adjustment: Manual verification of the identified pipeline information, denoted as I v =Validate(I); Manually adjust the pipeline with identification error, denoted as I f =Adjust(I v );

[0031] A4. 3D model construction and optimization:

[0032] Construct a 3D model: Construct a 3D model based on the extracted pipeline information, denoted as M 3D =Construct(I f ) to ensure that the geometry, size and position of the model are consistent with the actual pipeline;

[0033] Model optimization: simplify the model structure and reduce the number of polygons, denoted as M o =Optimize(M 3D ) to improve the rendering speed and interactive performance of the model;

[0034] A5. Export and integration:

[0035] Export model: Export the constructed 3D model to a format suitable for subsequent analysis and application, denoted as F=Export(M o );

[0036] Integration and verification: Integrate the exported 3D model into the BIM platform or AR device to verify the accuracy, completeness and consistency of the model, denoted as V(F).

[0037] Preferably, the working contents of the AR difference display and error identification module specifically include:

[0038] Import models into AR devices: Import both the drawing modeling and the actual modeling model into AR devices or AR-enabled applications in appropriate formats;

[0039] Differentiated display: Display two models simultaneously on an AR device using different colors, transparencies, or line styles; this can be achieved by setting the material properties of the models;

[0040] Error identification: Calculate the spatial distance between the corresponding points or line segments in the drawing modeling and the actual modeling. Assume that the position offsets of the spatial distance on the three coordinate axes of X, Y, and Z are Δx, Δy, and Δz respectively:

[0041] Δx=x2-x1;

[0042] Δy=y2-y1;

[0043] Δz=z2-z1;

[0044] Calculate the spatial error distance d using the Euclidean distance formula:

[0045] ;

[0046] Among them, (x1, y1, z1) and (x2, y2, z2) are the coordinates of the corresponding points in the drawing modeling and the actual modeling respectively;

[0047] Error threshold setting: define an error threshold to determine whether the difference between two models constitutes an error; the error threshold is based on the prediction of the regression model trained with historical data, and the error threshold is automatically adjusted according to project requirements and accuracy standards;

[0048] Error marking: Use visual markers and voice or text prompts on the AR device to mark errors that exceed the threshold;

[0049] Error marking: For points or line segments that exceed the error threshold, they are marked using visual markers and voice or text prompts on the AR device.

[0050] Preferably, the step of predicting the error threshold of the regression model based on historical data training specifically includes:

[0051] B1. Data collection and preprocessing:

[0052] Data collection: Collect data on error thresholds, installation accuracy, pipeline types, materials, and environmental conditions from multiple completed projects, and record the actual error thresholds used in each project and the corresponding installation results;

[0053] Data preprocessing: clean the collected data, remove incomplete or abnormal data points, and normalize the data to ensure that different features have the same weight in model training;

[0054] B2. Feature selection and model training:

[0055] Feature selection: Select features related to the error threshold setting from the collected data, including pipeline type, material, environmental conditions, and installation accuracy requirements, and use these features as input variables for the regression model;

[0056] Model training: Select linear regression as the model and use the preprocessed data to train the regression model. The goal is to predict the appropriate error threshold for a given feature. Then use the mean square error, root mean square error, or R² score to evaluate the performance of the model.

[0057] B3. The regression model formula is expressed as:

[0058] T=β0+β1·F1+β2·F2+...+β p ·F p ;

[0059] Among them, T is the prediction error threshold, β0 is the intercept term, (β1, β2, ..., β p ) are regression coefficients, (F1, F2, ..., F p ) are the selected features;

[0060] In the model training phase, the regression coefficients (β0, β1, ..., β p );

[0061] B4. Real-time application and feedback:

[0062] In new projects, the trained regression model is used to predict the appropriate error threshold according to the project characteristics and accuracy requirements, and the predicted error threshold is applied to the error identification process on the AR device to guide construction personnel in pipeline installation;

[0063] After the project is completed, data on the actual installation accuracy and the error threshold used are collected and used to update the regression model.

[0064] Preferably, after the spatial error distance d is calculated and obtained, an error estimation method based on the least squares method is further introduced to improve the error identification accuracy, and the formula is expressed as:

[0065] ;

[0066] Among them, E represents the sum of squares of errors, that is, the objective function;

[0067] p is a parameter vector, representing the transformation parameters to be solved;

[0068] d j represents the jth actual data point, that is, the three-dimensional coordinates of the pipeline point obtained during the actual scanning and modeling process;

[0069] M is a transformation matrix, which represents the transformation relationship from the drawing model to the actual model, including rotation, scaling and translation transformations, and is used to transform points in the drawing model into the actual coordinate system;

[0070] p j It represents the position that the jth point in the drawing model should reach after a certain transformation. The transformation is defined by the matrix M and the vector t. By finding the optimal transformation parameters M and t, the error between the transformed point and the actual point is minimized.

[0071] t is a translation vector used to adjust the overall position of the drawing model to align with the actual model;

[0072] n represents the number of data points, i.e., the total number of pipeline points obtained during the actual scanning and modeling process;

[0073] ‖·‖ represents the Euclidean norm of the vector, that is, the length or modulus of the vector, which is used to calculate the error between the actual point and the transformed point.

[0074] Preferably, the specific work contents of the correction suggestion and real-time guidance module include:

[0075] Generate correction suggestions: Generate correction suggestions based on the identified errors using algorithms or manual methods, including adjusting the position, height, angle and connection relationship of the pipeline;

[0076] Real-time guidance and correction: Correction suggestions are displayed on the AR device, and the display is updated in real time to reflect the corrected pipeline status; construction personnel refer to the instructions on the AR device to make adjustments.

[0077] Preferably, the step of generating correction suggestions specifically includes:

[0078] C1 Error statistics and analysis: After analyzing the 3D error of each point cloud data on the 3D coordinates, further extend and calculate the height error and angle error of a certain section of pipeline installation, determine the connection relationship error, classify and quantify the errors, and define the size and importance of each error;

[0079] C2 Select the correction method: Select the appropriate correction method according to the type and size of the error; for smaller errors, use manual adjustment; for complex or larger errors, use genetic algorithms or particle swarm optimization algorithms for correction;

[0080] C3 parameter setting: according to the actual situation of the pipeline, set the correction parameters, including position offset Δx, Δy, Δz, height adjustment Δh, angle adjustment Δθ;

[0081] C4 generates correction suggestions: Generates specific correction suggestions based on parameter values, including adjusting the position, height, angle or connection relationship of the pipeline; and displays the correction suggestions to the user in the AR device in graphical or textual form.

[0082] Preferably, the step of adjusting the pipeline in C4 specifically includes:

[0083] Position adjustment: Use Δx, Δy, and Δz to represent the position offset of the pipeline in three-dimensional space; calculate the position offset of each pipeline based on the results of the error analysis and apply it to the actual position adjustment of the pipeline:

[0084] (x′, y′, z′)=(x+Δx, y+Δy, z+Δz);

[0085] Among them, (x, y, z) is the original position coordinate, (x', y', z') is the adjusted position coordinate;

[0086] Height adjustment: Use Δh to represent the height adjustment of the pipeline; calculate the height adjustment amount based on the size and direction of the height error and apply it to the actual height adjustment of the pipeline:

[0087] h′=h+Δh;

[0088] Where h is the original height and h' is the adjusted height;

[0089] Angle adjustment: Δθ is used to represent the angle adjustment of the pipeline, where θ is the inclination angle or rotation angle of the pipeline. The angle adjustment is calculated based on the size and direction of the angle error and applied to the actual angle adjustment of the pipeline:

[0090] θ′=θ+Δθ;

[0091] Among them, θ is the original angle, and θ' is the adjusted angle;

[0092] Use trigonometric functions to calculate the angle direction between the coordinates of the two points of the adjusted pipeline;

[0093] Connection relationship adjustment: Redesign the connection method and sequence of pipelines based on connection relationship errors.

[0094] The present invention provides a BIM and AR-based electromechanical pipeline installation monitoring system. Compared with the prior art, it has the following beneficial effects:

[0095] 1. This BIM and AR-based electromechanical pipeline installation monitoring system achieves high-precision actual pipeline data collection through three-dimensional scanning and data preprocessing, providing a solid foundation for subsequent analysis. The introduction of the drawing three-dimensional modeling module makes the comparison between the design and the actual installation more intuitive. The application of AR difference display and error identification module not only improves the accuracy of error identification, but also predicts the error threshold through the regression model to effectively evaluate the severity of the error. The correction suggestion and real-time guidance module can provide correction suggestions in real time, guide construction personnel to operate, and improve construction efficiency and quality. Finally, the recording and feedback improvement module continues to accumulate data, optimizes the regression model, forms a virtuous circle, and continuously improves the monitoring and guidance capabilities of the system.

[0096] 2. The BIM and AR-based electromechanical pipeline installation monitoring system uses 3D scanning technology to capture the actual spatial characteristics of electromechanical pipelines with high precision. Through pre-processing steps such as denoising, registration and splicing, accurate and reliable 3D scanning data is generated, providing a solid foundation for subsequent modeling and analysis. Secondly, the drawing 3D modeling module uses BIM software and point cloud data generation technology to achieve accurate comparison between design drawings and actual installation conditions, improving modeling efficiency and accuracy. In particular, the application of technologies such as automated or semi-automated mesh generation and pipeline identification based on machine learning has further improved the automation and accuracy of modeling. Finally, the export and integration steps ensure the availability and accuracy of the 3D model in subsequent analysis and application, providing strong support for construction monitoring and guidance. These improvements have jointly promoted the advancement of electromechanical pipeline installation monitoring technology and improved construction quality and efficiency.

[0097] 3. The BIM and AR-based electromechanical pipeline installation monitoring system uses AR devices to simultaneously display drawing modeling and actual modeling, achieving intuitive difference display and facilitating construction personnel to quickly identify differences. Secondly, the introduction of spatial error distance calculation and regression models based on historical data training can accurately identify errors and predict appropriate error thresholds, thereby improving the accuracy and intelligence level of error identification. In addition, the error estimation method based on the least squares method further improves the accuracy of error identification and ensures the accuracy of pipeline installation. These improvements not only improve construction efficiency and quality, but also reduce rework costs caused by errors. At the same time, by continuously collecting actual installation data and updating the regression model, the system can continuously learn and optimize to adapt to more types of projects and higher precision requirements, promoting the continuous advancement of electromechanical pipeline installation technology.

[0098] 4. This BIM and AR-based electromechanical pipeline installation monitoring system accurately classifies and quantifies errors through error statistics and analysis, providing a reliable basis for the generation of correction suggestions. Secondly, according to the type and size of the error, the appropriate correction method is selected, and specific correction parameters are set to generate targeted correction suggestions, which improves the accuracy and efficiency of the correction. In addition, the real-time guidance correction function enables construction personnel to obtain correction suggestions instantly and make adjustments according to the instructions on the AR device, realizing the visualization and intelligence of the construction process. These improvements not only improve the accuracy and quality of electromechanical pipeline installation, but also reduce the difficulty and cost of construction, bringing more efficient and accurate solutions to the construction industry and promoting the innovation and development of construction technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0099] Figure 1 It is a system principle block diagram of the present invention;

[0100] Figure 2 A schematic flow chart of the steps for actually modeling the present invention;

[0101] Figure 3 A schematic diagram of the steps of predicting the error threshold of the regression model of the present invention;

[0102] Figure 4 Schematic flow chart of the steps for generating correction suggestions for the present invention. DETAILED DESCRIPTION

[0103] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0104] The present invention provides the following four technical solutions:

[0105] Figure 1 The first embodiment is shown: a BIM and AR-based electromechanical pipeline installation monitoring system, comprising:

[0106] Drawing 3D modeling module, used to create 3D drawing models of electromechanical pipelines based on design drawings;

[0107] The actual pipeline 3D modeling module uses the AR device shooting function to capture images of the actual pipeline during the actual installation process, processes the images to extract point cloud data of the actual pipeline, generates 3D scanning data through the pre-processing steps of denoising, registration, and splicing, and creates a model of the actually installed pipeline based on the 3D scanning data;

[0108] AR difference display and error identification module, used to import 3D drawing modeling and actual modeling into AR devices, distinguish and display the two models, calculate the spatial distance to identify errors, optimize the error calculation method to improve accuracy, and establish a regression model based on historical data training to predict threshold comparison errors and judge the severity of errors; at the same time, it provides error marking and correction suggestion display functions;

[0109] The correction suggestion and real-time guidance module classifies and quantifies the identified errors, defines the size and importance of each error, generates correction suggestions and displays them in real time on the AR device to guide construction personnel to make adjustments;

[0110] The recording and feedback improvement module records the pipeline spatial feature data before and after correction and the error data identified during the correction process, and records the correction suggestion data as historical data to train and improve the regression model.

[0111] Through 3D scanning and data preprocessing, high-precision actual pipeline data collection is achieved, providing a solid foundation for subsequent analysis. The introduction of the drawing 3D modeling module makes the comparison between the design and the actual installation more intuitive. The application of AR difference display and error identification module not only improves the accuracy of error identification, but also predicts the error threshold through the regression model to effectively evaluate the severity of the error. The correction suggestion and real-time guidance module can provide correction suggestions in real time, guide construction personnel to operate, and improve construction efficiency and quality. Finally, the recording and feedback improvement module continues to accumulate data, optimize the regression model, form a virtuous circle, and continuously improve the monitoring and guidance capabilities of the system.

[0112] Figure 2A second implementation is shown, which is mainly different from the first implementation in that: the work content of the drawing 3D modeling module specifically includes: creating a 3D drawing model of the electromechanical pipeline using BIM software according to the design drawing, wherein the parameters include defining the type, size, location and connection relationship of the pipeline;

[0113] The work content of the actual pipeline 3D modeling module specifically includes:

[0114] Point cloud data extraction: Select an AR device with a depth sensor to obtain depth information through other structured light or stereo vision methods, extract depth data from the image and generate point cloud data;

[0115] Data preprocessing:

[0116] Denoising: removing unwanted noise from scan data;

[0117] Registration: Use the ICP algorithm to align images from multiple perspectives into a common coordinate system;

[0118] Stitching: Use the point cloud stitching algorithm to merge the registered multi-view point cloud data into a complete 3D model data;

[0119] Actual modeling: import the preprocessed 3D model data into the modeling software, generate polygonal meshes or surfaces from the point cloud data, and create pipeline models based on the meshes or surfaces.

[0120] The specific steps of actual modeling include:

[0121] A1. Point cloud data import and preparation:

[0122] Import point cloud data: Use modeling software (such as AutoCAD Civil 3D, SketchUp) to import pre-processed point cloud data, denoted as P = {p1, p2, ..., p n}, where p i Represents the three-dimensional coordinates of the i-th point;

[0123] Data preparation: Clean the point cloud data and remove duplicate points and isolated points, denoted as P c =C(P); and according to the project requirements, the point cloud data is segmented and recorded as P s =S(P c );

[0124] A2. Automatic or semi-automatic mesh generation:

[0125] Automatic mesh generation: Use the built-in meshing tool in the software to convert point cloud data into polygonal meshes, denoted as M=Mesh(P s), set the mesh parameters including resolution r and smoothness s to ensure that the generated mesh is both accurate and meets the project requirements;

[0126] Semi-automatic adjustment: Use editing tools to manually adjust the mesh shape of the automatically generated mesh, denoted as M a =Adjust(M); make it more accurately reflect the shape and position of the actual pipeline;

[0127] A3. Pipeline information extraction and identification:

[0128] Pipeline recognition based on machine learning: The pipeline recognition algorithm based on machine learning is applied to process the grid data, denoted as I=Recognize(M a ), where I represents the extracted pipeline information;

[0129] When training the model, the labeled pipeline data is used as the training set, denoted as D = {(a1, b1), (a2, b2), ..., (a m , b m )}, where a k represents the input feature, b k Indicates the marked pipeline information;

[0130] Manual verification and adjustment: Manual verification of the identified pipeline information, denoted as I v =Validate(I); Manually adjust the pipeline with identification error, denoted as I f =Adjust(I v );

[0131] A4. 3D model construction and optimization:

[0132] Construct a 3D model: Construct a 3D model based on the extracted pipeline information, denoted as M 3D =Construct(I f ) to ensure that the geometry, size and position of the model are consistent with the actual pipeline;

[0133] Model optimization: simplify the model structure and reduce the number of polygons, denoted as M o =Optimize(M 3D ) to improve the rendering speed and interactive performance of the model;

[0134] A5. Export and integration:

[0135] Export model: Export the constructed 3D model to a format suitable for subsequent analysis and application, denoted as F=Export(M o ), where F represents the exported file format (such as OBJ, FBX, DWG, etc.);

[0136] Integration and verification: Integrate the exported 3D model into the BIM platform or AR device to verify the accuracy, completeness and consistency of the model, denoted as V(F); ensure that it is consistent with the actual pipeline installation situation.

[0137] 3D scanning technology can capture the actual spatial characteristics of electromechanical pipelines with high precision, and generate accurate and reliable 3D scanning data through pre-processing steps such as denoising, registration and splicing, providing a solid foundation for subsequent modeling and analysis. Secondly, the drawing 3D modeling module uses BIM software and point cloud data generation technology to achieve accurate comparison between design drawings and actual installation conditions, improving modeling efficiency and accuracy. In particular, the application of technologies such as automated or semi-automated mesh generation and pipeline identification based on machine learning has further improved the automation and accuracy of modeling. Finally, the export and integration steps ensure the availability and accuracy of the 3D model in subsequent analysis and application, providing strong support for construction monitoring and guidance. These improvements have jointly promoted the advancement of electromechanical pipeline installation monitoring technology and improved construction quality and efficiency.

[0138] Figure 3 The third embodiment is shown, and the main difference from the first embodiment is that the working content of the AR difference display and error identification module specifically includes:

[0139] Import models into AR devices: Import both the drawing model and the actual model into AR devices or AR-supported applications in appropriate formats (such as OBJ, FBX, etc.);

[0140] Differentiated display: Display two models simultaneously on an AR device using different colors, transparencies, or line styles; this can be achieved by setting the material properties of the models;

[0141] Error identification: Calculate the spatial distance between the corresponding points or line segments in the drawing modeling and the actual modeling. Assume that the position offsets of the spatial distance on the three coordinate axes of X, Y, and Z are Δx, Δy, and Δz respectively:

[0142] Δx=x2-x1;

[0143] Δy=y2-y1;

[0144] Δz=z2-z1;

[0145] Calculate the spatial error distance d using the Euclidean distance formula:

[0146] ;

[0147] Among them, (x1, y1, z1) and (x2, y2, z2) are the coordinates of the corresponding points in the drawing modeling and the actual modeling respectively;

[0148] Error threshold setting: define an error threshold to determine whether the difference between two models constitutes an error; the error threshold is based on the prediction of the regression model trained with historical data, and the error threshold is automatically adjusted according to project requirements and accuracy standards;

[0149] Error marking: Use visual markers and voice or text prompts on the AR device to mark errors that exceed the threshold;

[0150] Error marking: For points or line segments that exceed the error threshold, they are marked using visual markers (such as highlighting, arrow pointing, etc.) and voice or text prompts on the AR device.

[0151] The steps of predicting the error threshold of the regression model based on historical data training specifically include:

[0152] B1. Data collection and preprocessing:

[0153] Data collection: Collect data on error thresholds, installation accuracy, pipeline types, materials, and environmental conditions from multiple completed projects, and record the error thresholds actually used in each project and the corresponding installation results (such as whether the accuracy requirements are met, whether rework is required, etc.);

[0154] Data preprocessing: clean the collected data, remove incomplete or abnormal data points, and normalize the data to ensure that different features have the same weight in model training;

[0155] B2. Feature selection and model training:

[0156] Feature selection: Select features related to the error threshold setting from the collected data, including pipeline type, material, environmental conditions (such as temperature, humidity), and installation accuracy requirements, and use these features as input variables for the regression model;

[0157] Model training: Select linear regression as the model and use the preprocessed data to train the regression model. The goal is to predict the appropriate error threshold for a given feature. Then use the mean square error (MSE), root mean square error (RMSE), or R² score to evaluate the performance of the model.

[0158] B3. The regression model formula is expressed as:

[0159] T=β0+β1·F1+β2·F2+...+β p ·F p ;

[0160] Among them, T is the prediction error threshold, β0 is the intercept term, (β1, β2, ..., β p ) are regression coefficients, (F1, F2, ..., F p) are the selected features;

[0161] In the model training phase, the regression coefficients (β0, β1, ..., β p );

[0162] Techniques such as cross-validation can be used to optimize the parameters of the model, such as choosing the best polynomial degree, the depth of the decision tree, or the number of trees in the random forest;

[0163] B4. Real-time application and feedback:

[0164] In new projects, the trained regression model is used to predict the appropriate error threshold according to the project characteristics and accuracy requirements, and the predicted error threshold is applied to the error identification process on the AR device to guide construction personnel in pipeline installation;

[0165] After the project is completed, data collected on the actual installation accuracy and the error threshold used is used to update the regression model to improve the model's predictive accuracy for future projects.

[0166] After the spatial error distance d is calculated and obtained, the error estimation method based on the least squares method is further introduced to improve the error identification accuracy. The formula is expressed as:

[0167] ;

[0168] Among them, E represents the sum of squares of errors, that is, the objective function;

[0169] p is a parameter vector, representing the transformation parameters to be solved;

[0170] d j represents the jth actual data point, that is, the three-dimensional coordinates of the pipeline point obtained during the actual scanning and modeling process;

[0171] M is a transformation matrix, which represents the transformation relationship from the drawing model to the actual model, including rotation, scaling and translation transformations, and is used to transform points in the drawing model into the actual coordinate system;

[0172] p j It represents the position that the jth point in the drawing model should reach after a certain transformation. The transformation is defined by the matrix M and the vector t. By finding the optimal transformation parameters M and t, the error between the transformed point and the actual point is minimized.

[0173] t is a translation vector used to adjust the overall position of the drawing model to align with the actual model;

[0174] n represents the number of data points, i.e., the total number of pipeline points obtained during the actual scanning and modeling process;

[0175] ‖·‖ represents the Euclidean norm of the vector, that is, the length or modulus of the vector, which is used to calculate the error between the actual point and the transformed point.

[0176] The goal of the least squares method is to minimize the sum of the squares of the distances between all actual points and the transformed points by adjusting the transformation parameters M and t; this usually involves taking the derivative of the objective function E and setting the derivative equal to zero to find the optimal transformation parameters;

[0177] By simultaneously displaying the drawing modeling and the actual modeling through AR devices, an intuitive difference display is achieved, which is convenient for construction personnel to quickly identify the difference. Secondly, the introduction of spatial error distance calculation and regression model based on historical data training can accurately identify errors and predict appropriate error thresholds, improving the accuracy and intelligence level of error identification. In addition, the error estimation method based on least squares further improves the accuracy of error identification and ensures the accuracy of pipeline installation. These improvements not only improve construction efficiency and quality, but also reduce the rework costs caused by errors. At the same time, by continuously collecting actual installation data and updating the regression model, the system can continuously learn and optimize to adapt to more types of projects and higher precision requirements, promoting the continuous advancement of electromechanical pipeline installation technology.

[0178] Figure 4 The fourth embodiment is shown, which is mainly different from the first embodiment in that the specific work contents of the correction suggestion and real-time guidance module include:

[0179] Generate correction suggestions: Generate correction suggestions based on the identified errors using algorithms or manual methods, including adjusting the position, height, angle and connection relationship of the pipeline;

[0180] Real-time guidance and correction: Correction suggestions are displayed on the AR device, and the display is updated in real time to reflect the corrected pipeline status; construction personnel refer to the instructions on the AR device to make adjustments.

[0181] The steps to generate correction suggestions include:

[0182] C1 Error statistics and analysis: After analyzing the 3D error of each point cloud data on the 3D coordinates, further extend and calculate the height error and angle error of a certain section of pipeline installation, determine the connection relationship error, classify and quantify the errors, and define the size and importance of each error;

[0183] C2 Select the correction method: Select the appropriate correction method according to the type and size of the error; for smaller errors, use manual adjustment; for complex or larger errors, use algorithms (such as genetic algorithms, particle swarm optimization) for correction;

[0184] C3 parameter setting: according to the actual situation of the pipeline, set the correction parameters, including position offset Δx, Δy, Δz, height adjustment Δh, angle adjustment Δθ;

[0185] C4 generates correction suggestions: Generates specific correction suggestions based on parameter values, including adjusting the position, height, angle or connection relationship of the pipeline; and displays the correction suggestions to the user in a graphical or textual form in the AR device for the user to understand and implement.

[0186] The steps for adjusting the pipeline in C4 include:

[0187] Position adjustment: Use Δx, Δy, and Δz to represent the position offset of the pipeline in three-dimensional space; calculate the position offset of each pipeline based on the results of the error analysis and apply it to the actual position adjustment of the pipeline:

[0188] (x′, y′, z′)=(x+Δx, y+Δy, z+Δz);

[0189] Among them, (x, y, z) is the original position coordinate, (x', y', z') is the adjusted position coordinate;

[0190] Height adjustment: Use Δh to represent the height adjustment of the pipeline; calculate the height adjustment amount based on the size and direction of the height error and apply it to the actual height adjustment of the pipeline:

[0191] h′=h+Δh;

[0192] Among them, h is the original height, h' is the adjusted height;

[0193] Angle adjustment: Δθ is used to represent the angle adjustment of the pipeline, where θ is the inclination angle or rotation angle of the pipeline. The angle adjustment is calculated based on the size and direction of the angle error and applied to the actual angle adjustment of the pipeline:

[0194] θ′=θ+Δθ;

[0195] Among them, θ is the original angle, and θ' is the adjusted angle;

[0196] Use trigonometric functions (such as sin and cos) to calculate the angle direction between the coordinates of the two points of the adjusted pipeline;

[0197] Connection relationship adjustment: Redesign the connection method and sequence of pipelines according to the connection relationship errors; ensure that the adjusted pipeline connection relationship meets the design requirements while maintaining the fluidity and functionality of the system.

[0198] Through error statistics and analysis, errors are accurately classified and quantified, providing a reliable basis for the generation of correction suggestions. Secondly, according to the type and size of the error, the appropriate correction method is selected, and specific correction parameters are set to generate targeted correction suggestions, which improves the accuracy and efficiency of the correction. In addition, the real-time guidance correction function enables construction personnel to obtain correction suggestions instantly and make adjustments according to the instructions on the AR device, realizing the visualization and intelligence of the construction process. These improvements not only improve the accuracy and quality of mechanical and electrical pipeline installation, but also reduce the difficulty and cost of construction, bringing more efficient and accurate solutions to the construction industry and promoting the innovation and development of construction technology.

[0199] Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0200] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0201] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A BIM and AR-based electromechanical pipeline installation monitoring system, characterized in that: include: Drawing 3D modeling module, used to create 3D drawing models of electromechanical pipelines based on design drawings; The actual pipeline 3D modeling module uses the AR device shooting function to capture images of the actual pipeline during the actual installation process, processes the images to extract point cloud data of the actual pipeline, generates 3D scanning data through the pre-processing steps of denoising, registration, and splicing, and creates a model of the actually installed pipeline based on the 3D scanning data; AR difference display and error recognition module, used to import 3D drawing modeling and actual modeling into AR devices, distinguish and display the two models, calculate the spatial distance to identify errors, optimize the error calculation method to improve accuracy, and establish a regression model based on historical data training to predict threshold comparison errors and judge the severity of errors; It also provides error marking and correction suggestion display functions; The correction suggestion and real-time guidance module classifies and quantifies the identified errors, defines the size and importance of each error, generates correction suggestions and displays them in real time on the AR device to guide construction personnel to make adjustments; The recording and feedback improvement module records the pipeline spatial feature data before and after correction and the error data identified during the correction process, and records the correction suggestion data as historical data to train and improve the regression model.

2. According to claim 1, a BIM and AR-based electromechanical pipeline installation monitoring system is characterized by: The work content of the drawing three-dimensional modeling module specifically includes: according to the design drawings, using BIM software to create a three-dimensional drawing model of the electromechanical pipeline, wherein the parameters include defining the type, size, location and connection relationship of the pipeline.

3. According to the BIM and AR-based electromechanical pipeline installation monitoring system of claim 1, it is characterized by: The working content of the actual pipeline 3D modeling module specifically includes: Point cloud data extraction: Select an AR device with a depth sensor to obtain depth information through other structured light or stereo vision methods, extract depth data from the image and generate point cloud data; Data preprocessing: Denoising: removing unwanted noise from scan data; Registration: Use the ICP algorithm to align images from multiple perspectives into a common coordinate system; Stitching: Use the point cloud stitching algorithm to merge the registered multi-view point cloud data into a complete 3D model data; Actual modeling: import the preprocessed 3D model data into the modeling software, generate polygonal meshes or surfaces from the point cloud data, and create pipeline models based on the meshes or surfaces.

4. According to claim 3, a BIM and AR-based electromechanical pipeline installation monitoring system is characterized by: The specific steps of the actual modeling include: A1. Point cloud data import and preparation: Import point cloud data: Use modeling software to import pre-processed point cloud data, denoted as P = {p1, p2, ..., p n }, where p i Represents the three-dimensional coordinates of the i-th point; Data preparation: Clean the point cloud data and remove duplicate points and isolated points, denoted as P c =C(P); and according to the project requirements, the point cloud data is segmented and recorded as P s =S(P c ); A2. Automatic or semi-automatic mesh generation: Automatic mesh generation: Use the built-in meshing tool in the software to convert point cloud data into polygonal meshes, denoted as M=Mesh(P s ), set the grid parameters including resolution r and smoothness s; Semi-automatic adjustment: Use editing tools to manually adjust the mesh shape of the automatically generated mesh, denoted as M a =Adjust(M); A3. Pipeline information extraction and identification: Pipeline recognition based on machine learning: The pipeline recognition algorithm based on machine learning is applied to process the grid data, denoted as I=Recognize(M a ), where I represents the extracted pipeline information; When training the model, the labeled pipeline data is used as the training set, denoted as D = {(a1, b1), (a2, b2), ..., (a m , b m )}, where a k represents the input feature, b k Indicates the marked pipeline information; Manual verification and adjustment: Manual verification of the identified pipeline information, denoted as I v =Validate(I); Manually adjust the pipeline with identification error, denoted as I f =Adjust(I v ); A4. 3D model construction and optimization: Construct a 3D model: Construct a 3D model based on the extracted pipeline information, denoted as M 3D =Construct(I f ) to ensure that the geometry, size and position of the model are consistent with the actual pipeline; Model optimization: simplify the model structure and reduce the number of polygons, denoted as M o =Optimize(M 3D ) to improve the rendering speed and interactive performance of the model; A5. Export and integration: Export model: Export the constructed 3D model to a format suitable for subsequent analysis and application, denoted as F=Export(M o ); Integration and verification: Integrate the exported 3D model into the BIM platform or AR device to verify the accuracy, completeness and consistency of the model, denoted as V(F).

5. According to the BIM and AR-based electromechanical pipeline installation monitoring system of claim 1, it is characterized by: The work contents of the AR difference display and error identification module specifically include: Import models into AR devices: Import both the drawing modeling and the actual modeling model into AR devices or AR-enabled applications in appropriate formats; Differentiated display: Display two models simultaneously on an AR device using different colors, transparency, or line styles; Error identification: Calculate the spatial distance between the corresponding points or line segments in the drawing modeling and the actual modeling. Assume that the position offsets of the spatial distance on the three coordinate axes of X, Y, and Z are Δx, Δy, and Δz respectively: Δx=x2-x1; Δy=y2-y1; Δz=z2-z1; Calculate the spatial error distance d using the Euclidean distance formula: ； Among them, (x1, y1, z1) and (x2, y2, z2) are the coordinates of the corresponding points in the drawing modeling and the actual modeling respectively; Error threshold setting: define an error threshold to determine whether the difference between two models constitutes an error; the error threshold is based on the prediction of the regression model trained with historical data, and the error threshold is automatically adjusted according to project requirements and accuracy standards; Error marking: Use visual markers and voice or text prompts on the AR device to mark errors that exceed the threshold; Error marking: For points or line segments that exceed the error threshold, they are marked using visual markers and voice or text prompts on the AR device.

6. The BIM and AR-based electromechanical pipeline installation monitoring system according to claim 5 is characterized by: The steps of predicting the error threshold of the regression model based on historical data training specifically include: B1. Data collection and preprocessing: Data collection: Collect data on error thresholds, installation accuracy, pipeline types, materials, and environmental conditions from multiple completed projects, and record the actual error thresholds used in each project and the corresponding installation results; Data preprocessing: clean the collected data, remove incomplete or abnormal data points, and normalize the data to ensure that different features have the same weight in model training; B2. Feature selection and model training: Feature selection: Select features related to the error threshold setting from the collected data, including pipeline type, material, environmental conditions, and installation accuracy requirements, and use these features as input variables for the regression model; Model training: Select linear regression as the model and use the preprocessed data to train the regression model. The goal is to predict the appropriate error threshold for a given feature. Then use the mean square error, root mean square error, or R² score to evaluate the performance of the model. B3. The regression model formula is expressed as: T=β0+β1·F1+β2·F2+...+β p ·F p ; Among them, T is the prediction error threshold, β0 is the intercept term, (β1, β2, ..., β p ) are regression coefficients, (F1, F2, ..., F p ) are the selected features; In the model training phase, the regression coefficients (β0, β1, ..., β p ); B4. Real-time application and feedback: In new projects, the trained regression model is used to predict the appropriate error threshold according to the project characteristics and accuracy requirements, and the predicted error threshold is applied to the error identification process on the AR device to guide construction personnel in pipeline installation; After the project is completed, data on the actual installation accuracy and the error threshold used are collected and used to update the regression model.

7. The BIM and AR-based electromechanical pipeline installation monitoring system according to claim 5 is characterized by: After the spatial error distance d is calculated and obtained, an error estimation method based on the least squares method is further introduced to improve the error identification accuracy. The formula is expressed as: ； Among them, E represents the sum of squares of errors, that is, the objective function; p is a parameter vector, representing the transformation parameters to be solved; d j represents the jth actual data point, that is, the three-dimensional coordinates of the pipeline point obtained during the actual scanning and modeling process; M is a transformation matrix, which represents the transformation relationship from the drawing model to the actual model, including rotation, scaling and translation transformations, and is used to transform points in the drawing model into the actual coordinate system; p j It represents the position that the jth point in the drawing model should reach after a certain transformation. The transformation is defined by the matrix M and the vector t. By finding the optimal transformation parameters M and t, the error between the transformed point and the actual point is minimized. t is a translation vector used to adjust the overall position of the drawing model to align with the actual model; n represents the number of data points, i.e., the total number of pipeline points obtained during the actual scanning and modeling process; ‖·‖ represents the Euclidean norm of the vector, that is, the length or modulus of the vector, which is used to calculate the error between the actual point and the transformed point.

8. The BIM and AR-based electromechanical pipeline installation monitoring system according to claim 1 is characterized by: The specific work contents of the correction suggestion and real-time guidance module include: Generate correction suggestions: Generate correction suggestions based on the identified errors using algorithms or manual methods, including adjusting the position, height, angle and connection relationship of the pipeline; Real-time guidance and correction: Correction suggestions are displayed on the AR device, and the display is updated in real time to reflect the corrected pipeline status; construction personnel refer to the instructions on the AR device to make adjustments.

9. The BIM and AR-based electromechanical pipeline installation monitoring system according to claim 8, characterized in that: The step of generating correction suggestions specifically includes: C1 Error statistics and analysis: After analyzing the 3D error of each point cloud data on the 3D coordinates, further extend and calculate the height error and angle error of a certain section of pipeline installation, determine the connection relationship error, classify and quantify the errors, and define the size and importance of each error; C2 Select the correction method: Select the appropriate correction method according to the type and size of the error; for smaller errors, use manual adjustment; for complex or larger errors, use genetic algorithms or particle swarm optimization algorithms for correction; C3 parameter setting: according to the actual situation of the pipeline, set the correction parameters, including position offset Δx, Δy, Δz, height adjustment Δh, angle adjustment Δθ; C4 generates correction suggestions: Generates specific correction suggestions based on parameter values, including adjusting the position, height, angle or connection relationship of the pipeline; and displays the correction suggestions to the user in the AR device in graphical or textual form.

10. The BIM and AR-based electromechanical pipeline installation monitoring system according to claim 9, characterized in that: The steps for adjusting the pipeline in C4 include: Position adjustment: Use Δx, Δy, and Δz to represent the position offset of the pipeline in three-dimensional space; calculate the position offset of each pipeline based on the results of the error analysis and apply it to the actual position adjustment of the pipeline: (x′, y′, z′)=(x+Δx, y+Δy, z+Δz); Among them, (x, y, z) is the original position coordinate, (x', y', z') is the adjusted position coordinate; Height adjustment: Use Δh to represent the height adjustment of the pipeline; calculate the height adjustment amount based on the size and direction of the height error and apply it to the actual height adjustment of the pipeline: h′=h+Δh; Where h is the original height and h' is the adjusted height; Angle adjustment: Δθ is used to represent the angle adjustment of the pipeline, where θ is the inclination angle or rotation angle of the pipeline. The angle adjustment is calculated based on the size and direction of the angle error and applied to the actual angle adjustment of the pipeline: θ′=θ+Δθ; Among them, θ is the original angle, and θ' is the adjusted angle; Use trigonometric functions to calculate the angle direction between the coordinates of the two points of the adjusted pipeline; Connection relationship adjustment: Redesign the connection method and sequence of pipelines based on connection relationship errors.

Images