A geometric registration method and system based on BIM model and AR real scene

Abstract

The present invention relates to the field of highway engineering information technology, and in particular to a geometric alignment method and system based on a BIM model and an AR real scene. The present invention comprises the following steps: acquiring multi-source data of a construction site, generating an AR real scene of the construction site based on the multi-source data; fusing the multi-source data to obtain precise parameters of construction equipment; solving the spatial transformation matrix of the construction site by using precise parameters of the construction equipment, and obtaining the initial alignment result of the BIM model and the AR real scene by combining the engineering characteristics and the spatial transformation matrix; introducing a dynamic error compensation mechanism, and correcting the initial alignment result by using the dynamic error compensation mechanism to obtain a local alignment result; constructing a global and local coordinate system mapping optimization model, and using the global and local coordinate system mapping optimization model to map the local alignment result to the global coordinate system to obtain a global alignment result. The present invention provides a reliable spatial data foundation for highway engineering quality inspection, construction guidance, and defect identification.

Description

Technical Field

[0001] The present invention relates to the field of highway engineering information technology, and in particular to a geometric registration method and system based on a BIM model and an AR real scene. Background Art

[0002] BIM technology has become a crucial tool for the full lifecycle management of highway projects. Its capabilities, such as 3D visual modeling, collision detection, and quantity calculations, significantly improve design rationality and construction efficiency. However, during the construction phase, real-time matching of BIM models with actual on-site scenes remains challenging. Traditional methods require construction workers to visualize the space using 2D drawings or static BIM models. This can lead to construction errors due to misunderstandings and difficulty adapting to the dynamic changes in complex construction environments, resulting in significant spatial deviations between the model and the actual scene. While existing AR (augmented reality) technology can overlay virtual models onto the real scene, its registration accuracy is limited by sensor errors, environmental interference, and dynamic construction conditions, making it difficult to meet the precision required for quality inspections. Furthermore, existing collaborative applications of AR and BIM are primarily focused on indoor or simple-structured scenes. Geometric registration is less effective for large, linearly distributed, and geologically variable scenarios such as highway projects. Therefore, finding a method and system that can quickly and accurately perform geometric registration between BIM models and AR scenes, suitable for such large, linearly distributed, and geologically variable scenarios, is an urgent need. Summary of the Invention

[0003] In view of the shortcomings of existing methods and the needs of practical applications, in order to solve the problem of quickly and accurately realizing the geometric registration of BIM models and AR real scenes. On the one hand, the present invention provides a geometric registration method based on BIM models and AR real scenes, comprising the following steps: acquiring multi-source data of the construction site, generating an AR real scene of the construction site according to the multi-source data; fusing the multi-source data to obtain precise parameters of construction equipment; solving the spatial transformation matrix of the construction site by the precise parameters of the construction equipment, combining the engineering characteristics and the spatial transformation matrix to obtain the initial registration result of the BIM model and the AR real scene; introducing a dynamic error compensation mechanism, correcting the initial registration result by the dynamic error compensation mechanism, and obtaining a local registration result; constructing a global and local coordinate system mapping optimization model, and using the global and local coordinate system mapping optimization model to map the local registration result to the global coordinate system to obtain a global registration result.

[0004] The present invention integrates multi-source sensor data, constructs a dynamic spatial reference framework, introduces adaptive feature matching and error compensation mechanisms, and optimizes the mapping relationship between the local coordinate system and the global geographic information system in combination with the linear distribution characteristics of highway projects. It supports real-time dynamic calibration during construction, effectively reduces the impact of environmental interference, and achieves high-precision geometric alignment between BIM models and AR real scenes of construction sites. It provides a reliable spatial data foundation for quality inspection, construction guidance, and defect identification in construction scenarios such as highway projects that are large in scale, linearly distributed, and have variable geological conditions.

[0005] Optionally, fusing the multi-source data to obtain precise parameters of construction equipment includes the following steps:

[0006] The predicted parameters of the construction equipment are obtained based on the monitoring data of the inertial measurement unit; the prediction error of the predicted parameters is corrected by the Kalman gain to obtain the precise parameters of the construction equipment.

[0007] Optionally, the predicted parameters of the construction equipment are obtained based on the monitoring data of the inertial measurement unit, satisfying the following formula:

[0008] ,in, Indicates Always on time k The predicted value of the construction equipment system status at each moment, express The three-dimensional position of the construction equipment at any moment, express The three-dimensional speed of the construction equipment at any moment, Indicates the time difference, Represents the rotation matrix corresponding to the quaternion, express The posture of construction equipment at all times, express k The acceleration of time, represents the acceleration due to gravity, represents quaternion multiplication, represents the mapping from angular velocity to quaternion increments, express k The present invention achieves the unification and real-time update of the construction scene spatial reference by fusing multi-source sensor data, and provides a stable spatial transformation relationship for geometric alignment.

[0009] Optionally, the combining of the engineering characteristics and the spatial transformation matrix to obtain an initial registration result of the BIM model and the AR real scene comprises the following steps:

[0010] Extract key geometric features from the BIM model; and perform preliminary matching between the BIM model and the AR real scene based on the key geometric features.

[0011] Optionally, the initial matching of the BIM model and the AR real scene according to the key geometric features satisfies the following formula:

[0012] ,in, represents the minimization of the objective function, represents the rotation matrix, represents the translation vector, Represents the number of all matching point pairs, Indicates the The weight factor of the curvature similarity of the point pair satisfies: , Indicates the BIM model The curvature value of the feature point, Indicates AR real scene The curvature value of the matching points, Represents the weight decay speed adjustment parameter, Represents the first Feature points, Indicates AR real scene Matching points. Based on the spatial characteristics of the linear structure of highway engineering, the present invention extracts key geometric features from the BIM model and performs multi-scale matching with the actual point cloud of the construction site. At the same time, combined with dynamic disturbance monitoring and prediction, it realizes adaptive correction of the registration parameters, which is conducive to improving the accuracy of the present invention.

[0013] Optionally, the introducing of a dynamic error compensation mechanism and correcting the initial registration result by the dynamic error compensation mechanism to obtain a local registration result comprises the following steps:

[0014] Based on the monitoring data of the inertial measurement unit, the instantaneous displacement of the construction equipment is obtained; a future displacement trend prediction model of the construction equipment is constructed, and the posture increment is obtained by combining the future displacement trend prediction model, the instantaneous displacement and the monitoring data; the transformation matrix is ​​adjusted according to the posture increment, and the local alignment result is obtained using the adjusted transformation matrix.

[0015] Optionally, the transformation matrix is ​​adjusted according to the posture increment to satisfy the following formula:

[0016] ,in, represents the adjusted transformation matrix, represents the transformation matrix before adjustment, represents the 3×3 identity matrix, represents the rotation vector, Represents a rotation vector The corresponding antisymmetric matrix is, Represents the translation vector.

[0017] The present invention achieves error compensation through real-time monitoring and prediction mechanisms, eliminating dynamic interference in construction scenarios such as mechanical vibration and temporary facility displacement.

[0018] Optionally, the global and local coordinate system mapping optimization model is constructed to satisfy the following formula:

[0019] ,in, represents the minimization of the objective function, Represents the set of rotation, translation and scaling parameters to be optimized, Indicates the number of preset control point groups, Represents the coordinate transformation function, which maps the local coordinate system points to the global coordinate system. represents the three-dimensional coordinates of the i-th control point in the local construction coordinate system, represents the three-dimensional coordinates of the i-th control point in the global geographic coordinate system, Represents the square of the Euclidean distance, which is used to quantify the deviation between the local point and the global point after transformation.

[0020] Optionally, the local registration result is mapped to the global coordinate system using the global and local coordinate system mapping optimization model to obtain a global registration result that satisfies the following formula:

[0021]

[0022] in, represents the local coordinates, represents the inverse transformation matrix, Represents all coordinates, Indicates the The coordinates of the origin of the local coordinate system of a segment in the global coordinate system. By constructing a bidirectional mapping model and iteratively optimizing the transformation parameters, this invention eliminates the cumulative error of segment registration, solves the problem of nonlinear mapping deviation between the global geographic coordinate system and the local construction coordinate system in large-scale highway projects, and achieves geometric alignment consistency of the entire scene.

[0023] On the second aspect, in order to be able to efficiently execute the geometric alignment method based on BIM model and AR real scene provided by the present invention, the present invention also provides a geometric alignment system based on BIM model and AR real scene, including a processor, an input device, an output device and a memory, wherein the processor, input device, output device and memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the geometric alignment method based on BIM model and AR real scene as described in the first aspect of the present invention. The geometric alignment system based on BIM model and AR real scene of the present invention has a compact structure and stable performance, and can stably execute the geometric alignment method based on BIM model and AR real scene provided by the present invention, further improving the overall applicability and practical application capabilities of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A flow chart of a geometric registration method based on a BIM model and an AR real scene provided by an embodiment of the present invention;

[0025] Figure 2 A framework diagram of a geometric alignment system based on BIM models and AR real scenes provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0026] Specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are for illustrative purposes only and are not intended to limit the present invention. In the following description, numerous specific details are set forth to provide a thorough understanding of the present invention. However, it will be apparent to one of ordinary skill in the art that these specific details are not necessarily required to practice the present invention. In other instances, well-known circuits, software, or methods are not specifically described to avoid obscuring the present invention.

[0027] Throughout this specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in connection with the embodiment or example is included in at least one embodiment of the present invention. Therefore, appearances of the phrases "in one embodiment," "in an embodiment," "an example," or "an example" in various places throughout this specification are not necessarily all referring to the same embodiment or example. Furthermore, the particular features, structures, or characteristics may be combined in any suitable combinations and / or subcombinations in one or more embodiments or examples. Furthermore, those of ordinary skill in the art will appreciate that the figures provided herein are for illustrative purposes only and are not necessarily drawn to scale.

[0028] See also Figure 1In order to solve the problem of quickly and accurately realizing the geometric registration of BIM models and AR real scenes, the present invention provides a geometric registration method based on BIM models and AR real scenes, such as Figure 1 As shown, in one embodiment, the method includes the following steps:

[0029] S1. Acquire multi-source data of the construction site, and generate an AR real scene of the construction site based on the multi-source data.

[0030] In the embodiment, three-dimensional point clouds, real-scene models, and dynamic pose data of the construction site are collected synchronously through lidar, drone aerial survey, and inertial measurement unit (IMU), and multi-source data are denoised, coordinate normalized, and timestamp aligned to generate a standardized input data set.

[0031] Specifically, LiDAR is used to obtain high-density point cloud data of the construction site, with a scanning frequency of ≥20Hz and a point cloud resolution of ≤5cm; the drone is equipped with a multispectral camera and an RTK-GNSS module to generate an AR real scene of the construction site with a positioning accuracy of ±2cm; the inertial measurement unit (IMU) collects angular velocity, acceleration and posture data of the construction equipment at a frequency of 100Hz.

[0032] In some other embodiments, when acquiring multi-source data of a construction site, data preprocessing is also required, including point cloud denoising, coordinate normalization, and timestamp alignment.

[0033] Point cloud denoising refers to the removal of abnormal points in the point cloud based on statistical outlier removal (SOR). The average distance between the i-th point in the point cloud and its neighboring points k is calculated, and a threshold is set to remove noise points that deviate from the mean by more than 3 standard deviations.

[0034] , , elimination conditions: ,in, is the average Euclidean distance between the i-th point in the point cloud and the k points in its neighborhood, and N is the total number of point clouds.

[0035] Coordinate normalization refers to the uniform conversion of the BIM model (WGS84 coordinate system), point cloud (local coordinate system) and IMU data (equipment coordinate system) to the construction global coordinate system (CGCS2000).

[0036] Timestamp alignment refers to the use of PTP (Precision Time Protocol) to synchronize multi-sensor data with a time deviation of ≤1ms, ensuring data timing consistency.

[0037] Data preprocessing eliminates the heterogeneity of multi-source data through denoising, coordinate unification and time series synchronization, providing accurate, low-noise standardized input for subsequent alignment.

[0038] S2. Integrate the multi-source data to obtain accurate parameters of the construction equipment.

[0039] In an embodiment, the step S2 of fusing the multi-source data to obtain accurate parameters of the construction equipment includes the following steps:

[0040] S21. Obtain predicted parameters of the construction equipment based on the monitoring data of the inertial measurement unit.

[0041] First, define the global geographic coordinate system, local construction coordinate system and equipment movement coordinate system.

[0042] The global geographic coordinate system (GIS) refers to the integration of BIM model design parameters (road centerline coordinates, elevation data) based on the CGCS2000 coordinate system;

[0043] The local construction coordinate system is based on the highway pile number segment, defines the tangent direction along the road centerline as the x-axis, the normal direction as the y-axis, and the direction perpendicular to the ground as the z-axis, forming a linear reference system;

[0044] The equipment movement coordinate system refers to the IMU sensor bound to the construction machinery, with the origin being the center of mass of the equipment and the x-axis pointing in the direction of the equipment's movement.

[0045] Next, a piecewise linear mapping relationship is established to extract the highway centerline parametric equation from the BIM model. The arc length s is used as the independent variable, and the three-dimensional coordinates of the centerline are expressed as:

[0046] , where s is the arc length value corresponding to the pile number;

[0047] The construction scene is divided into several pile number intervals, and the mapping relationship between the local coordinate system and the global coordinate system in each interval is achieved through the homogeneous transformation matrix describe: ,in, is the rotation matrix, is the position of the origin of the local coordinate system of segment i in the global coordinate system.

[0048] Then, short-term state estimation is performed based on the monitoring data of the inertial measurement unit.

[0049] Construction equipment is constantly in motion during construction, and the IMU measures acceleration and angular velocity data 100 times per second. This data is used to predict the "next step" state of the equipment, and its state content requires predicting the three-dimensional position of the construction equipment. , three-dimensional velocity , and posture (The posture is represented by quaternions, which are similar to three-dimensional rotation angles).

[0050] The status of construction equipment can be expressed as:

[0051]

[0052] Furthermore, the acceleration measured by the inertial measurement unit and angular velocity .

[0053] Furthermore, position change is calculated by multiplying the current velocity by the time interval to estimate the next position. For example, if the device moves eastward at 1 m / s, its position will increase by 0.1 m eastward after 0.1 seconds. Speed ​​change is calculated by multiplying the acceleration measured by the IMU by the time interval to estimate the change in velocity. For example, if the acceleration is 2 m / s², the velocity will increase by 0.2 m / s after 0.1 seconds. Attitude change is calculated by calculating the change in quaternion attitude using the IMU's angular velocity (how fast the device rotates) and updating the device's orientation. If the device turns left, the attitude parameters are adjusted accordingly.

[0054] Specifically, the predicted parameters of the construction equipment are obtained based on the monitoring data of the inertial measurement unit, and satisfy the following formula:

[0055] ,in, Indicates Always on time k The predicted value of the construction equipment system status at each moment, express The three-dimensional position of the construction equipment at any moment, express The three-dimensional speed of the construction equipment at any moment, Indicates the time difference, Represents the rotation matrix corresponding to the quaternion, which is used to convert the acceleration from the device coordinate system to the global coordinate system. express The posture of construction equipment at all times, express k The acceleration of time, represents the acceleration due to gravity, represents quaternion multiplication, represents the mapping from angular velocity to quaternion increments, express k The angular velocity at the moment, .

[0056] Furthermore, the mapping from angular velocity to quaternion increment satisfies:

[0057]

[0058] Furthermore, the observation data includes the position after LiDAR matching and GNSS positioning , the observation equation is: ,in, represents zero-mean Gaussian observation noise.

[0059] Furthermore, the state transfer Jacobian matrix satisfies: , and is the partial derivative of position and velocity with respect to quaternion attitude, which needs to be calculated through the Lie algebra perturbation model.

[0060] in, A skew-symmetric matrix representing the angular velocity.

[0061] The observation Jacobian matrix satisfies:

[0062] , process noise covariance, satisfies: ,in and are the noise standard deviations of the accelerometer and gyroscope, respectively.

[0063] S22. Correct the prediction error of the prediction parameter by using the Kalman gain to obtain the precise parameters of the construction equipment.

[0064] The prediction results may be biased due to IMU noise and need to be corrected using actual measurement data from LiDAR and GNSS:

[0065] First, a lidar system scans the surrounding environment, matches feature points in the BIM model, and calculates the device's actual position (with an accuracy of approximately 5cm). Second, the drone provides high-precision global positioning (horizontal accuracy of ±2cm), but this may be temporarily ineffective due to weather or obstructions. Finally, it directly provides the device's real-time orientation (such as pitch and roll angles).

[0066] The predicted position and attitude are compared with the actual measurements from LiDAR and GNSS, yielding the difference (e.g., a predicted position is 1.2 meters east, while the actual measurement is 1.15 meters, resulting in a residual error of 0.05 meters). The Kalman gain is used to determine whether to trust the predicted or observed values ​​more. For example, if the GNSS signal is stable (with low error), the GNSS data is given more credence, and the prediction is significantly revised. If the LiDAR fails to match due to dust interference, its weight is reduced, and primary reliance is placed on the IMU prediction and GNSS data.

[0067] According to the residual and Kalman gain, the position, speed and attitude parameters of the construction equipment are adjusted to obtain a better estimated value, that is, the precise parameters of the construction equipment.

[0068] Furthermore, if GNSS suddenly fails, the extended Kalman filter (EKF) automatically increases the weight of LiDAR and IMU to ensure positioning continuity. In response to the above abnormal situation, it then updates the transformation matrix 100 times per second to reflect device movement and posture changes in real time, providing a precise spatial reference for AR model overlay.

[0069] The Extended Kalman Filter uses the IMU to rapidly detect the device's motion trends. Combined with field measurement data from LiDAR and GNSS, it dynamically balances the errors between prediction and measurement, ultimately outputting reliable position, velocity, and attitude information. This process effectively addresses the insufficient accuracy of a single sensor in construction environments and establishes a real-time, adaptive spatial benchmark for geometric registration.

[0070] S3. Calculate the spatial transformation matrix of the construction site using the precise parameters of the construction equipment, and obtain an initial registration result of the BIM model and the AR real scene by combining the engineering characteristics and the spatial transformation matrix.

[0071] In an embodiment, the spatial transformation matrix of the construction site calculated by using the precise parameters of the construction equipment refers to a dynamic spatial reference framework based on the geographic coordinate system (GIS), the local construction coordinate system and the equipment movement coordinate system, and the spatial transformation matrix obtained by fusing the precise parameters of the construction equipment according to the extended Kalman filter.

[0072] Furthermore, the combining of the engineering characteristics and the spatial transformation matrix to obtain an initial registration result of the BIM model and the AR real scene includes the following steps:

[0073] S31. Extract key geometric features from the BIM model.

[0074] Specifically, the key geometric features include the highway centerline, slope profile, pile number marking points, and curvature-normal vector features of the slope profile.

[0075] Feature discrimination is enhanced by calculating the curvature C and normal vector n of the local feature points of the point cloud, where the curvature is determined by the ratio of the eigenvalues ​​of the covariance matrix, and the normal vector is solved by principal component analysis (PCA) to find the direction corresponding to the minimum eigenvalue.

[0076] S32: Preliminary matching is performed between the BIM model and the AR real scene according to the key geometric features.

[0077] The AR real-world point cloud is voxel-downsampled (voxel size 0.1m), and the curvature-normal vector features corresponding to the BIM are extracted. The BIM features are then pre-registered with the downsampled AR real-world point cloud using iterative closest point (ICP) method.

[0078] Furthermore, the BIM model and the AR real scene are initially matched according to the key geometric features, satisfying the following formula:

[0079] ,in, represents the minimization of the objective function, represents the rotation matrix, represents the translation vector, Represents the number of all matching point pairs, Indicates the The weight factor of the curvature similarity of the point pair satisfies: , Indicates the BIM model The curvature value of the feature point, Indicates the curvature value of the matching point in the AR scene. Represents the weight decay speed adjustment parameter, Represents the BIM model Feature points, Indicates AR real scene Matching points.

[0080] S4. Introduce a dynamic error compensation mechanism, and correct the initial registration result through the dynamic error compensation mechanism to obtain a local registration result.

[0081] After the initial matching of local features is completed, dynamic interference in the construction scene, such as mechanical vibration and displacement of temporary facilities, may still cause continuous deviation of the registration parameters, requiring error compensation through real-time monitoring and prediction mechanisms.

[0082] Specifically, the introduction of the dynamic error compensation mechanism and the correction of the initial registration result by the dynamic error compensation mechanism to obtain the local registration result include the following steps:

[0083] S41. Obtain instantaneous displacement of the construction equipment based on monitoring data from the inertial measurement unit.

[0084] The IMU collects the angular velocity ω and acceleration a of the construction equipment at a frequency of 100 Hz, separates the gravity component g through integration, and calculates the instantaneous displacement of the equipment. : , which reflects the posture deviation of the device caused by vibration or collision.

[0085] S42: Construct a future displacement trend prediction model for the construction equipment, and obtain a posture increment by combining the future displacement trend prediction model, the instantaneous displacement, and the monitoring data.

[0086] Specifically, the future displacement trend prediction model is an LSTM network model.

[0087] Input the pose sequence within 10 seconds of history , output the displacement prediction value for the next 3 seconds ,The network structure is 3-layer LSTM, with 64 nodes in the hidden layer, and the ,mean square error (MSE) loss function is used for training.

[0088] By weighted fusion of the prediction results and the IMU measured values, the pose increment is generated to meet the following requirements: ,in, represents the pose increment, represents the rotation vector, Represents the translation vector.

[0089] S43. Adjust the transformation matrix according to the posture increment, and obtain a local registration result using the adjusted transformation matrix.

[0090] Specifically, the transformation matrix is ​​adjusted according to the posture increment to satisfy the following formula: ,in, represents the adjusted transformation matrix, represents the transformation matrix before adjustment, represents the 3×3 identity matrix, represents the rotation vector, Represents a rotation vector The corresponding antisymmetric matrix is, Represents the translation vector.

[0091] Furthermore, the initial registration result obtained in the above steps is adjusted using the adjusted transformation matrix to obtain a local registration result.

[0092] In the embodiment, the displacement threshold of the construction equipment is set to 5 cm. If the equipment is subjected to continuous vibration within 1 second, resulting in a cumulative instantaneous displacement greater than the displacement threshold of 5 cm, the registration parameters need to be corrected. Otherwise, no correction is required.

[0093] S5. Construct a global and local coordinate system mapping optimization model, and use the global and local coordinate system mapping optimization model to map the local registration result to the global coordinate system to obtain a global registration result.

[0094] While dynamic error compensation ensures local registration stability, it is necessary to address the nonlinear mapping deviation between the global geographic coordinate system and the local construction coordinate system in large-scale highway projects to achieve geometric alignment consistency across the entire scene. By constructing an optimization model for mapping global and local coordinate systems and iteratively optimizing the conversion parameters, we can eliminate the accumulated errors in segmented registration and ensure registration accuracy.

[0095] Specifically, highway engineering presents linear distribution characteristics, and its local construction coordinate system is based on the pile number segmentation. The coordinate system of each segment is transformed by the homogeneous transformation matrix Associated with the Global Geographic Coordinate System (CGCS2000).

[0096] The constructed global and local coordinate system mapping optimization model satisfies the following formula: ,in, represents the minimization of the objective function, Represents the set of rotation, translation and scaling parameters to be optimized, Indicates the number of preset control point groups, Represents the coordinate transformation function, which maps the local coordinate system points to the global coordinate system. represents the three-dimensional coordinates of the i-th control point in the local construction coordinate system, represents the three-dimensional coordinates of the i-th control point in the global geographic coordinate system, Represents the square of the Euclidean distance, which is used to quantify the deviation between the local point and the global point after transformation.

[0097] Furthermore, the Levenberg-Marquardt algorithm is used to iteratively solve the problem. The algorithm dynamically adjusts the damping factor Balancing the convergence of gradient descent and Gauss-Newton methods: , where is the residual r parameter The Jacobian matrix of is the parameter increment, and the parameters are updated iteratively until the residual converges.

[0098] If the residual decreases, then ;

[0099] If the residual increases, .

[0100] Furthermore, the global and local coordinate system mapping optimization model is used to map the local registration result to the global coordinate system to obtain a global registration result, which satisfies the following formula:

[0101] in, represents the local coordinates, represents the inverse transformation matrix, Represents all coordinates, Indicates the The coordinates of the origin of the segment's local coordinate system in the global coordinate system. Due to construction errors and environmental factors, the actual coordinate transformation has nonlinear deviations, and the mapping parameters need to be optimized through control points.

[0102] Furthermore, after completing the mapping optimization of the global and local coordinate systems, the high-precision aligned BIM model and AR real scene need to be superimposed on the construction terminal, and interactive operations are used to provide feedback and guide on-site operations.

[0103] See also Figure 2 In an embodiment, in order to efficiently execute the geometric registration method based on BIM models and AR real scenes provided by the present invention, the present invention also provides a geometric registration system based on BIM models and AR real scenes, including: an input device, an output device, a processor, and a memory, wherein the input device, output device, processor, and memory are interconnected, and the memory contains program instructions, which are used for the steps of the geometric registration method based on BIM models and AR real scenes. The geometric registration system based on BIM models and AR real scenes of the present invention has a compact structure and stable performance, and can stably execute the geometric registration method based on BIM models and AR real scenes of the present invention, further improving the overall applicability and practical application capabilities of the present invention.

[0104] In an embodiment, the processor may be a central processing unit (CPU), which may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. The input device may be used to obtain data information. The output device may be used to output the results obtained by storing the program instructions contained in the computer program in the memory provided by the present invention. The memory may include a read-only memory and a random access memory, and provides instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory.

[0105] In one possible implementation, the memory may include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function, etc.; the data storage area may store data created during use. In addition, the memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include NVRAM. The memory stores an operating system and operating instructions, executable modules or data structures, or a subset thereof, or an extended set thereof, wherein the operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and processing hardware-based tasks.

[0106] An embodiment further provides a storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the above-mentioned geometric registration method based on the BIM model and the AR real scene are implemented.

[0107] The storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc., which can store program codes.

[0108] In summary, the present invention integrates multi-source sensor data, constructs a dynamic spatial reference framework, introduces adaptive feature matching and error compensation mechanisms, and optimizes the mapping relationship between the local coordinate system and the global geographic information system in combination with the linear distribution characteristics of highway engineering. It supports real-time dynamic calibration during construction, effectively reduces the impact of environmental interference, and achieves high-precision geometric alignment between the BIM model and the AR real scene of the construction site, providing a reliable spatial data foundation for quality inspection, construction guidance and defect identification in construction scenarios such as highway engineering, which have large scope, linear distribution and changeable geological conditions.

[0109] Therefore, the present invention effectively overcomes various shortcomings of the prior art and has high industrial utilization value.

[0110] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope described in the present invention.

Claims

1. A geometric registration method based on BIM model and AR real scene, characterized by: The following steps are involved: Acquire multi-source data of the construction site, and generate an AR real scene of the construction site based on the multi-source data; Fusing the multi-source data to obtain accurate parameters of construction equipment; The spatial transformation matrix of the construction site is calculated by using the precise parameters of the construction equipment, and the initial registration result of the BIM model and the AR real scene is obtained by combining the engineering characteristics and the spatial transformation matrix; Introducing a dynamic error compensation mechanism, and correcting the initial registration result through the dynamic error compensation mechanism to obtain a local registration result; Constructing a global and local coordinate system mapping optimization model, and using the global and local coordinate system mapping optimization model to map the local registration result to a global coordinate system to obtain a global registration result; The method of fusing the multi-source data to obtain accurate parameters of the construction equipment includes the following steps: Obtain predicted parameters of construction equipment based on monitoring data from the inertial measurement unit; Correcting the prediction error of the prediction parameter by using the Kalman gain to obtain the precise parameters of the construction equipment; The predicted parameters of the construction equipment are obtained based on the monitoring data of the inertial measurement unit, which satisfies the following formula: in, Indicates Always on time k The predicted value of the construction equipment system status at each moment, express The three-dimensional position of the construction equipment at any moment, express The three-dimensional speed of the construction equipment at any moment, Indicates the time difference, Represents the rotation matrix corresponding to the quaternion, express The posture of construction equipment at all times, express k The acceleration of time, represents the acceleration due to gravity, represents quaternion multiplication, represents the mapping from angular velocity to quaternion increments, express k Angular velocity at the moment; Combining the engineering characteristics and the spatial transformation matrix to obtain an initial registration result between the BIM model and the AR real scene includes the following steps: extracting key geometric features from the BIM model; Performing a preliminary matching between the BIM model and the AR real scene according to the key geometric features; The BIM model and the AR real scene are initially matched according to the key geometric features, satisfying the following formula: in, represents the minimization of the objective function, represents the rotation matrix, represents the translation vector, Represents the number of all matching point pairs, Indicates the The weight factor of the curvature similarity of the point pair satisfies: , Indicates the BIM model The curvature value of the feature point, Indicates AR real scene The curvature value of the matching points, Represents the weight decay speed adjustment parameter, Represents the first Feature points, Indicates AR real scene Matching points.

2. The geometric registration method based on BIM model and AR real scene according to claim 1 is characterized in that: The introducing of the dynamic error compensation mechanism and correcting the initial registration result by the dynamic error compensation mechanism to obtain the local registration result include the following steps: Based on the monitoring data of the inertial measurement unit, the instantaneous displacement of the construction equipment is obtained; Constructing a future displacement trend prediction model for the construction equipment, and obtaining a posture increment by combining the future displacement trend prediction model, the instantaneous displacement, and the monitoring data; The transformation matrix is ​​adjusted according to the posture increment, and the local registration result is obtained using the adjusted transformation matrix.

3. The geometric registration method based on BIM model and AR real scene according to claim 2 is characterized in that: The transformation matrix is ​​adjusted according to the posture increment to satisfy the following formula: in, represents the adjusted transformation matrix, represents the transformation matrix before adjustment, represents the 3×3 identity matrix, represents the rotation vector, Represents a rotation vector The corresponding antisymmetric matrix is, Represents the translation vector.

4. The geometric registration method based on BIM model and AR real scene according to claim 1 is characterized in that: The global and local coordinate system mapping optimization model is constructed to satisfy the following formula: in, represents the minimization of the objective function, Represents the set of rotation, translation and scaling parameters to be optimized, Indicates the number of preset control point groups, Represents the coordinate transformation function, which maps the local coordinate system points to the global coordinate system. represents the three-dimensional coordinates of the i-th control point in the local construction coordinate system, represents the three-dimensional coordinates of the i-th control point in the global geographic coordinate system, Represents the square of the Euclidean distance, which is used to quantify the deviation between the local point and the global point after transformation.

5. The geometric registration method based on BIM model and AR real scene according to claim 1 is characterized in that: The local registration result is mapped to the global coordinate system by using the global and local coordinate system mapping optimization model to obtain a global registration result that satisfies the following formula: in, represents the local coordinates, represents the inverse transformation matrix, Represents all coordinates, Indicates the The coordinates of the origin of the segment's local coordinate system in the global coordinate system.

6. A geometric registration system based on BIM model and AR real scene, characterized by: The geometric registration system based on BIM model and AR real scene includes: an input device, an output device, a processor, and a memory, wherein the input device, output device, processor, and memory are interconnected, and the memory includes program instructions, and the program instructions are used to execute the geometric registration method based on BIM model and AR real scene according to any one of claims 1 to 5.

Images