A concrete vibrating construction control system based on AR visual laser scanning

Abstract

The application belongs to the technical field of concrete construction quality control, and discloses a concrete vibrating construction control system based on AR visual laser scanning, which comprises a total controller, a three-dimensional laser scanner, a vibrating rod sensor group and an AR visual device which are respectively communicatively connected with the total controller; the three-dimensional laser scanner is used for collecting real-time site measured point clouds of a vibrating area and uploading the total controller during vibrating operation; the vibrating rod sensor group is used for collecting vibrating depth, vibrating frequency and vibrating time length and uploading the total controller; and the total controller comprises a BIM analysis module, a registration calculation module, a space-time coupling module and a judgment and early warning module. The application improves the construction efficiency and pouring quality of vibrating construction; meanwhile, through self-adaptive dimming, the early warning mark is clearly superimposed on the real scene picture under any lighting condition, so that the operator can accurately obtain quality control prompts at all times.

Description

Technical Field

[0001] This invention belongs to the field of concrete construction quality control technology, specifically relating to a concrete vibration construction control system based on AR visual laser scanning. Background Technology

[0002] Patent document CN120471500A discloses a building quality inspection management system, specifically a system based on BIM, point cloud acquisition, multi-sensor integration, and AR. However, this solution focuses on post-construction random inspections of component defects, rebar arrangement, and concrete strength, and is not suitable for process control during concrete pouring and vibration. In existing concrete construction site vibration control operations, the pouring zones and vibration standard parameter information planned in advance by BIM are mostly stored separately within the design terminal. They cannot be shared with the real-time point cloud data collected by laser scanning at the construction site and the operational data collected by the sensors attached to the vibrator. These three types of data belong to different independent systems and lack a unified spatial coordinate as a connecting link. It is difficult to accurately match the standard vibration points and zoning ranges marked on the design drawings with the actual operating positions of the vibrator during the construction process.

[0003] Most 3D laser point cloud registration methods used on construction sites employ fixed-weight ICP algorithms. The algorithm parameters do not change autonomously with the ambient lighting conditions. Under different working conditions, such as strong direct sunlight during the day and dim lighting on the shaded side, the point cloud registration calculation error fluctuates significantly, easily leading to misalignment between the baseline model and the measured points on site. Furthermore, during the concrete pouring stage, the dense reinforcing bars and formwork components frequently obstruct the scanning optical path, causing missing point cloud data in localized areas. Conventional equipment, after a brief signal interruption from the laser scanner, cannot rely on other auxiliary data to continue tracking the real-time position of the vibrator, further exacerbating the problem of inaccurate spatial positioning.

[0004] Currently, the mainstream data processing model sends all collected data, such as vibration depth, working frequency, and operation time, to the backend for processing before providing feedback. Construction personnel cannot directly view the original measured parameters on-site. Issues such as insufficient vibration depth, excessively short or long vibration time, vibration frequency deviation from design values, and missed vibration in local areas can mostly only be identified through sampling inspections after construction is completed. Potential problems cannot be detected in a timely manner during construction.

[0005] Under the traditional management model, even if post-construction inspections reveal that the vibration construction is substandard, on-site operators lack intuitive references for correction. There are no directional signs or correction routes marked on the actual field of vision. They can only rely on the workers' long-term accumulated construction experience to adjust the vibration position and operation status independently. It is difficult to avoid vibration quality defects caused by human error.

[0006] In summary, the existing vibration control scheme has problems such as data lag, poor visualization, and lack of real-time guidance on site. Summary of the Invention

[0007] The technical problem to be solved by the present invention is to provide a concrete vibration construction control system based on AR visual laser scanning, which addresses the shortcomings of the prior art.

[0008] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: A concrete vibration construction control system based on AR vision laser scanning includes a main controller, a three-dimensional laser scanner, a vibrator sensor group, and an AR vision device for communicating with the main controller respectively. The three-dimensional laser scanner is used to collect real-time on-site measured point cloud data of the vibration area during vibration operation and upload it to the main controller; The vibratory rod sensor group is used to collect three types of vibration working condition data in real time: vibration depth, vibration frequency and vibration duration, and upload them to the main controller. The main controller includes a BIM parsing module, a registration calculation module, a spatiotemporal coupling module, and a judgment and early warning module; The BIM parsing module is used to acquire BIM data to generate a three-dimensional reference model of the pouring area. Based on the quality control requirements for vibration construction, it divides the three-dimensional reference model into vibration grid zones and marks the vibration points corresponding to each grid. The registration and calculation module uses the ICP algorithm to dynamically register the measured point cloud on site with the three-dimensional reference model, constructing a unified virtual-real coordinate system. The spatiotemporal coupling module extracts the spatial coordinates of the vibrator and the vibration condition data within the registered virtual-real coordinate system and performs time-series synchronization binding, constructing an association relationship between spatial, temporal, and conditional integrated data. Based on this association relationship, the judgment and early warning module determines the vibration condition status according to a preset threshold group, generates a corresponding identifier, and sends it to the AR vision device. The main controller matches the three types of vibration condition data with corresponding vibration points based on the association relationship and forwards them to the AR vision device for real-world display. The AR vision device overlays operating parameters and warning signs onto the corresponding vibration points based on the correlation, and adaptively adjusts the brightness and transparency of the virtual information according to the on-site lighting.

[0009] Furthermore, the spatiotemporal coupling module establishes a unified time series by setting a minimum time unit, and each acquisition device samples synchronously according to a set sampling frequency to achieve time sequence alignment of multi-source data; at the same time, using the synchronization timestamp as an index, the spatial coordinate data of the vibrator registered at the same time is bound one by one with the vibration condition data, thus constructing the correlation relationship of integrated spatial, temporal and condition data.

[0010] Furthermore, the registration operation module generates the virtual and real coordinate systems through the following steps: 1) Select three non-collinear reference points in the pouring area, collect the coordinates of the reference points on site with the AR vision device, and compare them with the coordinates of the corresponding reference points in the three-dimensional reference model to complete the initial alignment; 2) Perform preprocessing and coordinate centering on the benchmark model point cloud and the field measured point cloud respectively; 3) The measured point cloud on site is mapped to the reference model coordinate system through coordinate transformation, and a virtual and real coordinate system is formed after accurate registration; During the registration process, if the local point cloud occlusion rate is ≥30%, it is determined that there is an abnormal point cloud occlusion, and neighborhood feature interpolation is used to fill the points; if the duration of a single laser interruption is ≤2s, it is determined that there is a short-term interruption abnormality of laser scanning, and the vibration position is predicted in real time by relying on the AR built-in inertial measurement unit.

[0011] Furthermore, the preprocessing process is as follows: first, Gaussian filtering is used to perform neighborhood weighted smoothing to remove outliers, and then pass-through filtering is used to remove redundant points.

[0012] Furthermore, a set of point clouds for a three-dimensional baseline model is defined. ,in, For the 3D reference point cloud set, the first A three-dimensional coordinate point, The total number of point clouds in the 3D baseline model. It is a three-dimensional real space; a collection of measured point clouds from the field. , The first point cloud set measured on site A three-dimensional coordinate point, This represents the total number of point clouds measured on-site. The transformation formula for mapping the measured point cloud to the coordinate system of the three-dimensional reference model is as follows: ,in, This is the rotation matrix of the point cloud measured on-site. The translation vector of the measured point cloud is given; iterative solution is used. The transformed set of on-site measured point clouds Point cloud set of 3D benchmark model Minimum overall deviation; Coordinate centering is achieved by calculating the centroids of two sets of point clouds respectively: The centroid expression for the point cloud of the 3D baseline model is: ; The expression for the centroid of the point cloud measured in the field is: ; The expression for the centered coordinates is: , .

[0013] Furthermore, the ICP algorithm is an improved version of ICP that introduces illumination-adaptive weights. The expression for the illumination-adaptive weights is:

[0014] in, For illumination-adaptive weights, The intensity of ambient light; Construct the weighted distance error function: ,in, The Euclidean L2 norm square operator is used to iteratively minimize the error function, solve for the optimal rotation matrix and translation vector, and perform adaptive lighting weighting. It switches in real time according to the lighting conditions, reducing the interference of environmental noise on registration accuracy; The ICP algorithm iteratively sets dual termination conditions: iteration stops when the mean square error of the iteration is ≤0.5mm or the number of iterations reaches 50.

[0015] Furthermore, the early warning module determines the vibration condition status based on a preset threshold group. The determination process is as follows: A red anomaly marker is generated when the insertion depth is less than the first depth threshold, and a depth correction guide line is sent to the AR. A green normal marker is generated when the insertion depth is greater than or equal to the second depth threshold. A yellow warning marker is generated when the first depth threshold is less than or equal to the second depth threshold. The first depth threshold is 40mm to 45mm, and the second depth threshold is 50mm to 55mm. A red abnormality marker is generated when the single-point vibration duration is less than the first duration threshold or greater than or equal to the fourth duration threshold; a yellow warning marker is generated when the first duration threshold is less than or equal to the second duration threshold or the third duration threshold is less than or equal to the fourth duration threshold; and a green normal marker is generated when the second duration threshold is less than or equal to the third duration threshold. The first duration threshold is 13s to 17s, the second duration threshold is 18s to 22s, the third duration threshold is 28s to 32s, and the fourth duration threshold is 33s to 37s. A red abnormality is determined when the vibration frequency is less than the first frequency threshold or greater than the third frequency threshold. A yellow abnormality is generated when the first frequency threshold is less than or equal to the second frequency threshold or when the second frequency threshold is less than or equal to the third frequency threshold. A green normality is generated when the vibration frequency equals the second duration threshold. The first frequency threshold is 185Hz to 195Hz, the second frequency threshold is 198Hz to 202Hz, and the third frequency threshold is 205Hz to 215Hz.

[0016] Furthermore, the process of generating the deep correction guide line is as follows: the real-time spatial coordinates of the vibrator are used as the starting point of the guide line, and the coordinates of the standard design vibration point in the three-dimensional reference model are used as the ending point of the guide line. The two points are connected to form the guide line and superimposed on the AR real scene.

[0017] Furthermore, the vibratory rod sensor group consists of a depth sensor, a vibration frequency sensor, and a timing module; the 3D laser scanner, AR vision device, and main controller use 5G and WiFi dual-mode communication to transmit data.

[0018] Compared with the prior art, the present invention has the following advantages: This invention utilizes a 3D laser scanner, a vibratory rod sensor group, and AR vision equipment, breaking the traditional model of separating monitoring and data viewing. Based on BIM, it pre-divides vibratory grid zones, unifying the virtual and real coordinate system to achieve integrated control of construction space and vibration parameters. The spatiotemporal coupling module matches the corresponding grid zones according to the vibratory rod coordinates, precisely binding spatial location with working condition data such as vibration depth, frequency, and duration, ensuring that each pouring point corresponds to complete vibration parameters. Raw sensor data is directly pushed to the AR real-world display, and the judgment and warning module generates hierarchical status indicators. Operators can intuitively grasp the vibration quality and adjust the insertion depth in real time, eliminating the traditional model of delayed background viewing and post-event rectification, effectively improving construction efficiency and pouring quality. Addressing issues such as large differences in lighting at construction sites, overexposure of virtual indicators under strong light, and glare in low-light environments, adaptive dimming technology ensures that warning indicators are clearly superimposed on the real-world image, enabling operators to stably obtain quality control prompts around the clock and avoiding the impact of light and dark imbalances on on-site judgment.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system application process according to an embodiment of the present invention. Detailed Implementation

[0021] Example of a concrete vibration construction control system based on AR visual laser scanning: To address the issue that existing monitoring data can only be accessed and viewed in the background, and that operators cannot visually determine the vibration compliance status and depth adjustment direction in real time, relying heavily on post-event review and rectification, resulting in insufficient overall construction efficiency and pouring quality control accuracy.

[0022] like Figure 1As shown, the concrete vibration construction control system based on AR visual laser scanning includes a main controller, a 3D laser scanner, a vibratory rod sensor group, and an AR vision device for communication connection with the main controller. The AR vision device uses helmet-mounted AR glasses. Preferably, a miniature 3D laser scanning module is integrated and installed in the frame of the AR vision glasses, synchronously collecting measured point clouds as the personnel move. The vibratory rod sensor group consists of a depth sensor, a vibration frequency sensor, and a timing module; the 3D laser scanner, AR vision device, and main controller use 5G and WiFi dual-mode communication to transmit data. This solves the problem of independent hardware and data isolation between the traditional BIM design end, on-site scanning end, and vibration sensing end, making it convenient for operators to follow the system's guidance and adjust construction methods in a timely manner.

[0023] The aforementioned 3D laser scanner is used to collect real-time measured point clouds of the vibration area during vibration operations and upload them to the main controller. By collecting dynamic real-time point clouds of the pouring face throughout the entire process, the system provides real-time feedback on the spatial topography of the construction site and the real-time spatial position of the vibrator, achieving dynamic spatial perception throughout the entire vibration process. The aforementioned vibrator sensor group is used to collect real-time data on three types of vibration conditions: vibration depth, vibration frequency, and vibration duration, and upload this data to the main controller. The main controller includes a BIM analysis module, a registration calculation module, a spatiotemporal coupling module, and a judgment and early warning module.

[0024] Specifically, the aforementioned BIM parsing module is used to acquire BIM data to generate a three-dimensional reference model of the pouring area. Based on the quality control requirements for vibration compaction construction, it divides the three-dimensional reference model into vibration grid zones and marks the vibration points corresponding to each grid. According to vibration compaction construction standards, the vibration grid is customized, solving the problem that traditional zoning sizes are uniform and difficult to adapt to all construction standards.

[0025] For example, BIM is used to fully model the large-volume concrete pouring area of ​​the nuclear island raft foundation, including the raft foundation pit edge line, steel mesh, pre-embedded anchors, and pre-embedded pipes to build a high-precision three-dimensional benchmark model of the nuclear island raft foundation.

[0026] Therefore, in accordance with the nuclear power plant civil construction specifications and the special construction plan for large-volume concrete, the designed vibration area is displayed in the BIM model. The large-area raft foundation is automatically divided into several vibration grid unit zones according to the 2m×2m vibration grid. According to the vibration construction quality control requirements, the vibration grid zones are divided in the three-dimensional reference model. The regular areas are divided into fixed-size standard grid units, and the special restricted areas are arranged with smaller grid units.

[0027] The aforementioned registration module uses the ICP algorithm to dynamically register the measured point cloud with the 3D reference model, constructing a unified virtual-real coordinate system. Specifically, the registration module generates the virtual-real coordinate system through the following steps: 1) Three non-collinear reference points are selected in the pouring area. The AR vision device collects the coordinates of the reference points on site and compares them with the coordinates of the corresponding reference points in the 3D reference model to complete the initial alignment. Utilizing the principle of three non-collinear spatial positioning, the virtual and real models are coarsely aligned, reducing the registration range of subsequent algorithm iterations and decreasing the computational load of the ICP algorithm iteration.

[0028] 2) Preprocessing and coordinate centering were performed on the baseline model point cloud and the field-measured point cloud, respectively. The preprocessing process was as follows: first, Gaussian filtering was used to perform neighborhood weighted smoothing to remove outliers, and then pass-through filtering was used to remove redundant points. Outliers refer to points formed by dust and light reflection; redundant points refer to irrelevant debris on the ground. This targeted removal of two types of invalid point clouds—dust reflection, light noise, and ground debris—precisely reduced point cloud noise interference caused by construction dust, construction waste, and light reflection.

[0029] 3) The measured point cloud on site is mapped to the reference model coordinate system through coordinate transformation, and a virtual and real coordinate system is formed after accurate registration; During registration, if the local point cloud occlusion rate is ≥30%, an anomaly is identified, and neighborhood feature interpolation is used to supplement the points. This ensures uninterrupted ICP registration operations and maintains the spatial positioning continuity of the vibrator even under complex occlusion conditions. If the duration of a single laser interruption is ≤2s, a short-term laser scanning interruption is identified, and the vibration position is predicted in real time using the AR's built-in inertial measurement unit (IMU). This compensates for the limitations of single-position laser scanning and improves the fault tolerance of positioning under complex construction site conditions.

[0030] Specifically, (1) the process of interpolating and supplementing points using neighborhood features is as follows: Let the coordinates of the newly generated supplementary point at the missing position be... Take the area around the missing point. Neighboring valid points Weighted interpolation is performed, and the interpolation calculation formula is as follows:

[0031] in, Supplementary coordinate points are generated by weighted interpolation for missing locations in the point cloud due to occlusion. The total number of valid neighborhood points selected. For the first The weighting coefficients corresponding to each effective neighboring point To cover the area around the missing section The coordinates of valid points in the neighborhood are used. The occluded and missing point cloud is completed by neighborhood weighted interpolation, which ensures that the subsequent ICP registration operation is continuous and uninterrupted, and improves the system's operational stability under the condition of material occlusion at the construction site.

[0032] (2) The process of calculating the spatial position of vibration in real time based on the built-in inertial measurement unit (IMU) of the AR vision device is as follows: the attitude transformation matrix is ​​solved by using the three-axis attitude angles output in real time by the IMU. Combined with the last valid coordinates before the laser interruption The formula for predicting the current spatial coordinates of the vibrator is as follows:

[0033] in, The real-time spatial coordinates of the vibratory rod are obtained from the IMU prediction. This is the attitude transformation matrix obtained from the IMU attitude angle calculation. This is the last set of valid vibratory rod coordinates acquired before the laser interruption. This prediction method maintains continuous registration operations, avoiding spatial coordinate drift. The system sequentially integrates illumination-adaptive weighted improved ICP iterative convergence control, point cloud occlusion neighborhood interpolation missing compensation, and IMU short-term disconnection position multi-compensation mechanism. Under multiple constraints, the overall point cloud registration error is low. Based on the linkage between historical valid coordinates and real-time attitude matrix, dynamic position tracking is achieved under short-term disconnection conditions.

[0034] Define a set of point clouds for a 3D baseline model ,in, For the 3D reference point cloud set, the first A three-dimensional coordinate point, The total number of point clouds in the 3D baseline model. It is a three-dimensional real space; a collection of measured point clouds from the field. , The first point cloud set measured on site A three-dimensional coordinate point, This represents the total number of point clouds measured on-site. The transformation formula for mapping the measured point cloud to the coordinate system of the three-dimensional reference model is as follows: ,in, This is the rotation matrix of the point cloud measured on-site. This is the translation vector of the point cloud measured on-site.

[0035] The coordinate centering process is achieved by calculating the centroids of two sets of point clouds: The centroid expression for the point cloud of the 3D baseline model is: ; The expression for the centroid of the point cloud measured in the field is: ; The expression for the centered coordinates is: , By centering the centroids of the two point clouds, the original position offsets of the two sets of point clouds are eliminated.

[0036] The ICP algorithm is an improved version of ICP that introduces illumination-adaptive weights. The expression for the illumination-adaptive weights is:

[0037] in, For illumination-adaptive weights, The ambient light intensity is used as the threshold. The algorithm iteratively adjusts the weighting coefficients based on a 500 Lux light intensity value to adapt to both strong and dim lighting conditions at the construction site.

[0038] Construct the weighted distance error function: ,in, The Euclidean L2 norm square operator is used to iteratively minimize the error function, solve for the optimal rotation matrix and translation vector, and perform adaptive lighting weighting. The system switches in real time with the illumination, reducing the interference of environmental noise on registration accuracy. Illumination weights are embedded in the error function to dynamically correct point matching error values ​​based on illumination.

[0039] The ICP algorithm iteratively sets dual termination conditions: iteration stops when the mean square error is ≤0.5mm or the number of iterations reaches 50. The iterative solution process is as follows: 1) Find the nearest point pair: For each field measurement point scanned by the 3D laser scanner In the point cloud of the three-dimensional baseline model Find the corresponding point with the minimum Euclidean distance in the middle: ,in, For on-site measurement points With the three-dimensional reference model points The Euclidean distance between them; using the Euclidean distance to match virtual and real points of the same origin, and locking in a one-to-one corresponding reference point pair.

[0040] 2) Construct the covariance matrix: ,in, The covariance matrix is ​​constructed for the centralized 3D benchmark model point cloud and the centralized field measured point cloud to characterize the spatial correlation between the two sets of centralized point clouds. Centralized on-site measurement points The transpose of .

[0041] 3) Singular Value Decomposition (SVD) pairs Perform SVD decomposition: Find the optimal rotation matrix ,in, , For the covariance matrix The two sets of orthogonal matrices obtained after singular value decomposition For matrix The transpose matrix is ​​obtained. SVD matrix decomposition is used to optimally solve the spatial rotation parameters. Compared with gradient iterative solution, SVD decomposition provides higher accuracy and faster convergence, making it suitable for the real-time dynamic registration computing power requirements of construction sites.

[0042] 4) Find the translation vector ; 5) Update point cloud , This represents the current iteration number of the ICP algorithm; No. The coordinates of the measured points in the next iteration To complete the updated coordinates of the field measurement points in this iteration; 6) Iteration termination condition: Calculate the overall mean square error after each iteration: Stop iteration when MSE ≤ 0.5 mm, or when the number of iterations ≥ 50. The mean square error of the overall matching between the transformed field measured point cloud and the 3D benchmark model point cloud is used as the criterion for stopping the ICP iteration.

[0043] The aforementioned spatiotemporal coupling module extracts and synchronously binds the spatial coordinates of the vibrator in the registered virtual-real coordinate system with the vibration condition data, constructing a correlation between spatial, temporal, and condition-integrated data. Specifically, the spatiotemporal coupling module establishes a unified time series using a set minimum time unit, and each acquisition device synchronously samples at a set sampling frequency to achieve temporal alignment of multi-source data. Simultaneously, using the synchronization timestamp as an index, it binds the registered spatial coordinate data of the vibrator with the vibration condition data at the same moment, constructing a correlation between spatial, temporal, and condition-integrated data. Acquisition devices refer to a 3D laser scanner, a vibrator sensor group, and an AR vision device. Constructing this correlation between spatial, temporal, and condition-integrated data provides unified and standardized input data for vibration condition threshold determination, AR real-world parameter visualization, and point cloud anomaly compensation. It enables traceability of the corresponding depth, frequency, and duration parameters for any vibration point throughout the entire time period, supporting real-time early warning and post-event quality traceability.

[0044] To ensure complete temporal consistency among multi-source data sources, including 3D laser scanning point clouds, AR images, and vibratory rod sensors, the system constructs a globally unified time reference and sets a minimum time unit. .

[0045] Define a globally unified time series: ,in, This is the sampling sequence number.

[0046] All acquisition devices strictly follow this time series for synchronous sampling: the 3D laser scanner and AR vision module are configured with different sampling frequencies. The sampling interval is consistent with the minimum time unit; various sensors mounted on the vibratory rod synchronously trigger sampling, with each timestamp... A unique set of spatial positions of vibrating rods and a complete set of working parameters are used to achieve strong coupling and binding in the temporal dimension.

[0047] In order to establish the correlation between spatial, temporal, and operational condition integrated data, the dataset is first defined, and the registered spatial coordinates are set as follows: The three-dimensional spatial coordinate expression of the vibrator after time registration is as follows: Based on the vibration grid unit, the spatial coordinates are partitioned and mapped to obtain the vibration partition number corresponding to the current position. Then the sequence of motion trajectories of the vibrator is: .

[0048] Same timestamp Below, the vibratory rod sensor simultaneously collects three raw working parameters: vibration depth. Vibration frequency Single-point cumulative vibration time .

[0049] Therefore, the mapping relationship expression is:

[0050] in, It is a spatiotemporal coupling mapping function. For the first The three-dimensional spatial coordinates of the vibrating rod at all times For the first The vibration bar is located in the vibration grid zone number at any given time. For the first The depth of the vibratory rod insertion at all times. For the first The vibration frequency of the vibrating rod at all times, For the first The cumulative vibration time at a single point at any given moment.

[0051] At the same time Using the three-dimensional spatial coordinates of the vibratory rod and its corresponding vibration zone as indexes, all sensor condition data at that moment are bound together, ultimately achieving the integration of multi-source data from the spatial domain, time domain, and operational condition parameters, providing a complete and standardized data source for real-time quality control and early warning via the AR interface and precise traceability of vibration quality.

[0052] To provide clearer operational guidance for workers, the judgment and early warning module, based on the aforementioned correlation, determines the vibration condition status according to a preset threshold group and generates corresponding indicators, which are then sent to the AR vision device. The preset threshold group is set according to the pouring plan. The bound vibration condition data is transmitted to the judgment and early warning module for comparison with the preset thresholds to generate abnormal warning indicators; simultaneously, it is pushed to the AR vision device for intuitive display. Operators can read the operational parameters in real time and independently judge the working condition, forming a dual quality control guarantee of automatic system early warning and manual experience verification. This changes the traditional mode of delayed background judgment and post-event sampling inspection, enabling real-time identification of potential problems such as under-vibration, over-vibration, insufficient depth, and excessive frequency on-site, and issuing corrective instructions immediately.

[0053] Specifically, the above-mentioned main controller classifies the working condition based on the insertion depth, single-point vibration duration, and vibration frequency according to a preset threshold group.

[0054] When the insertion depth is less than the first depth threshold, a red abnormality marker is generated and a depth correction guide line is sent to the AR system. When the insertion depth is greater than or equal to the second depth threshold, a green normal marker is generated. When the first depth threshold is less than or equal to the insertion depth but less than the second depth threshold, a yellow warning marker is generated. The first depth threshold is 40mm to 45mm, and the second depth threshold is 50mm to 55mm. In this embodiment, the first depth threshold is 45mm, and the second depth threshold is 50mm. The generation process of the depth correction guide line is as follows: the real-time spatial coordinates of the vibrator are used as the starting point of the guide line, and the coordinates of the standard design vibrating point in the three-dimensional reference model are used as the ending point of the guide line. The two points are connected to form the guide line, which is then superimposed on the AR real-world image.

[0055] A red abnormality marker is generated when the single-point vibration duration is less than the first duration threshold or greater than or equal to the fourth duration threshold. A yellow warning marker is generated when the first duration threshold is less than or equal to the second duration threshold or the third duration threshold is less than or equal to the fourth duration threshold. A green normal marker is generated when the second duration threshold is less than or equal to the third duration threshold. The first duration threshold is 13s to 17s, the second duration threshold is 18s to 22s, the third duration threshold is 28s to 32s, and the fourth duration threshold is 33s to 37s. In this embodiment, the first duration threshold is 15s, the second duration threshold is 20s, the third duration threshold is 30s, and the fourth duration threshold is 35s.

[0056] A red anomaly is identified when the vibration frequency is less than the first frequency threshold or greater than the third frequency threshold. A yellow anomaly is generated when the first frequency threshold is less than or equal to the second frequency threshold, or when the second frequency threshold is less than or equal to the third frequency threshold. A green normal status is generated when the vibration frequency equals the second duration threshold. The first frequency threshold is 185Hz–195Hz, the second frequency threshold is 198Hz–202Hz, and the third frequency threshold is 205Hz–215Hz. The thresholds can be determined based on the engineering requirements for concrete and the allowable vibration frequency of the selected vibrator. In this embodiment, the first frequency threshold is 190Hz, the second frequency threshold is 200Hz, and the third frequency threshold is 210Hz. Nuclear power engineering has high requirements for compaction, therefore, higher frequency thresholds are used.

[0057] Based on the aforementioned correlation, the central controller matches the three types of vibration condition data with the corresponding vibration points and forwards them to the AR vision device for real-world display. On-site personnel can directly read the original construction parameters, achieving dual quality control through system early warning and manual verification.

[0058] AR vision devices adaptively adjust the brightness and transparency of virtual information based on ambient lighting; the virtual information includes vibration data and other information displayed on the AR vision devices. The AR device also adapts to ambient lighting and dynamically adjusts the image quality parameters of virtual signs to avoid lighting issues that could cause workers to misread warning signs or misoperate, ensuring effective quality control throughout the day.

[0059] The process in this embodiment is as follows: Workers wear AR vision devices to perform concrete vibration operations. The vibrator sensor group collects three types of working condition data in real time: vibration depth, vibration frequency, and vibration duration. This data is then uploaded to the central controller via 5G and WiFi dual-mode communication. The central controller uses an improved ICP algorithm with adaptive weighting for illumination, combined with point cloud occlusion interpolation and laser short-interruption IMU position prediction mechanisms, to achieve dynamic registration between the measured point cloud on site and the BIM 3D benchmark model, constructing a unified virtual-real coordinate system. The spatiotemporal coupling module binds the vibration spatial coordinates and working condition data to form an integrated relationship. The judgment and early warning module determines the working status against a preset vibration threshold and generates a three-color graded label, which is then sent to the AR vision device. Based on this integrated relationship, the AR vision device overlays the vibration grid partitions, working condition parameters, and early warning labels onto the corresponding vibration points, and adaptively adjusts the brightness and transparency of the virtual information according to the on-site illumination. If the vibration working condition parameters deviate from the threshold, the AR interface displays a corresponding color-coded early warning label and a depth correction guide line, achieving real-time visual quality control throughout the entire vibration operation process.

[0060] Addressing the pain points of large-scale, layered, continuous pouring of large-volume concrete for nuclear island raft foundations, stringent quality control in critical areas, and the tendency for traditional manual vibration to result in missed vibration, under-vibration, and poor uniformity, this invention completely eliminates missed vibration and blank areas in the large-area vibration of the raft foundation. The abnormal rate of under-vibration and over-vibration is controlled within 2%, significantly improving the overall uniformity of concrete density and increasing the pass rate of physical testing to over 98%. Simultaneously, it enables visualization, quantification, and correction of the vibration process for key nuclear island structures, meeting the high standards, zero-defect, and full traceability quality control requirements of nuclear power engineering. This significantly reduces the costs of manual inspection and subsequent rework, and substantially improves the construction quality and standardization level of large-volume concrete pouring.

[0061] The above description is merely a preferred embodiment of the present invention and does not constitute any limitation on the present invention. Any simple modifications, alterations, or equivalent structural changes made to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A concrete vibration construction control system based on AR visual laser scanning, characterized in that: This includes a central controller, as well as a 3D laser scanner, a vibrating rod sensor group, and an AR vision device for communicating with the central controller. The three-dimensional laser scanner is used to collect real-time on-site measured point cloud data of the vibration area during vibration operation and upload it to the main controller; The vibratory rod sensor group is used to collect three types of vibration working condition data in real time: vibration depth, vibration frequency and vibration duration, and upload them to the main controller. The main controller includes a BIM parsing module, a registration calculation module, a spatiotemporal coupling module, and a judgment and early warning module; The BIM parsing module is used to acquire BIM data to generate a three-dimensional reference model of the pouring area. According to the quality control requirements of vibration construction, it divides the three-dimensional reference model into vibration grid zones and marks the vibration points corresponding to each grid. The registration calculation module uses the ICP algorithm to dynamically register the on-site measured point cloud with the three-dimensional reference model to construct a unified virtual and real coordinate system. The spatiotemporal coupling module extracts the spatial coordinates of the vibrator and the vibration condition data in the registered virtual and real coordinate system and performs time-series synchronous binding to construct the correlation relationship of integrated spatial, temporal, and condition data. The judgment and early warning module determines the vibration condition status based on the correlation relationship and according to the preset threshold group, generates the corresponding identifier and sends it to the AR vision device; the main controller matches the three types of vibration condition data with the corresponding vibration points based on the correlation relationship and forwards them to the AR vision device for real-scene display. The AR vision device overlays operating parameters and warning signs onto the corresponding vibration points based on the correlation, and adaptively adjusts the brightness and transparency of the virtual information according to the on-site lighting.

2. The concrete vibration construction control system based on AR visual laser scanning according to claim 1, characterized in that: The spatiotemporal coupling module establishes a unified time series by setting a minimum time unit, and each acquisition device samples synchronously according to a set sampling frequency to achieve time sequence alignment of multi-source data; at the same time, using the synchronization timestamp as an index, the spatial coordinate data of the vibrator registered at the same moment is bound one by one with the vibration condition data, thus constructing the correlation relationship of integrated spatial, temporal and condition data.

3. A concrete vibration construction control system based on AR visual laser scanning as described in claim 1, characterized in that: The registration calculation module generates the virtual and real coordinate systems through the following steps: 1) Select three non-collinear reference points in the pouring area, collect the coordinates of the reference points on site with the AR vision device, and compare them with the coordinates of the corresponding reference points in the three-dimensional reference model to complete the initial alignment; 2) Perform preprocessing and coordinate centering on the benchmark model point cloud and the field measured point cloud respectively; 3) The measured point cloud on site is mapped to the reference model coordinate system through coordinate transformation, and a virtual and real coordinate system is formed after accurate registration; During the registration process, if the local point cloud occlusion rate is ≥30%, it is determined that there is an abnormal point cloud occlusion, and neighborhood feature interpolation is used to fill the points; if the duration of a single laser interruption is ≤2s, it is determined that there is a short-term interruption abnormality of laser scanning, and the vibration position is predicted in real time by relying on the AR built-in inertial measurement unit.

4. A concrete vibration construction control system based on AR visual laser scanning according to claim 3, characterized in that: The preprocessing process is as follows: first, Gaussian filtering is used to perform neighborhood weighted smoothing to remove outliers, and then pass-through filtering is used to remove redundant points.

5. A concrete vibration construction control system based on AR visual laser scanning according to claim 3, characterized in that: Define a set of point clouds for a 3D baseline model ,in, For the 3D reference point cloud set, the first A three-dimensional coordinate point, The total number of point clouds in the 3D baseline model. It is a three-dimensional real space; a collection of measured point clouds from the field. , The first point cloud set measured on site A three-dimensional coordinate point, This represents the total number of point clouds measured on-site. The transformation formula for mapping the measured point cloud to the coordinate system of the three-dimensional reference model is as follows: ,in, This is the rotation matrix of the point cloud measured on-site. The translation vector of the measured point cloud is obtained through iterative solution. The transformed set of on-site measured point clouds Point cloud set of 3D benchmark model Minimum overall deviation; Coordinate centering is achieved by calculating the centroids of two sets of point clouds respectively: The centroid expression for the point cloud of the 3D baseline model is: ; The expression for the centroid of the point cloud measured in the field is: ; The expression for the centered coordinates is: , .

6. A concrete vibration construction control system based on AR visual laser scanning according to claim 5, characterized in that: The ICP algorithm described is an improved ICP that incorporates adaptive illumination weights. The expression for the adaptive illumination weights is as follows: in, For illumination-adaptive weights, The intensity of ambient light; Construct the weighted distance error function: ,in, The Euclidean L2 norm square operator is used to iteratively minimize the error function, solve for the optimal rotation matrix and translation vector, and perform adaptive lighting weighting. It switches in real time according to the lighting conditions, reducing the interference of environmental noise on registration accuracy; The ICP algorithm iteratively sets dual termination conditions: iteration stops when the mean square error of the iteration is ≤0.5mm or the number of iterations reaches 50.

7. A concrete vibration construction control system based on AR visual laser scanning according to claim 1, characterized in that: The judgment and early warning module determines the vibration condition status based on a preset threshold group. The judgment process is as follows: When the insertion depth is less than the first depth threshold, a red abnormality indicator is generated and a depth correction guide line is sent to AR. When the insertion depth is greater than or equal to the second depth threshold, a green normal indicator is generated. When the first depth threshold is less than or equal to the insertion depth and less than the second depth threshold, a yellow warning indicator is generated. The first depth threshold is 40mm to 45mm, and the second depth threshold is 50mm to 55mm. A red abnormality indicator is generated when the single-point vibration time is less than the first time threshold or greater than the fourth time threshold; a yellow warning indicator is generated when the first time threshold is less than the second time threshold or the third time threshold is less than the fourth time threshold; and a green normal indicator is generated when the second time threshold is less than the third time threshold. The first duration threshold is 13s to 17s, the second duration threshold is 18s to 22s, the third duration threshold is 28s to 32s, and the fourth duration threshold is 33s to 37s. A red abnormality is determined when the vibration frequency is less than the first frequency threshold or greater than the third frequency threshold. A yellow abnormality is generated when the first frequency threshold is less than or equal to the second frequency threshold or when the second frequency threshold is less than or equal to the third frequency threshold. A green normality is generated when the vibration frequency equals the second duration threshold. The first frequency threshold is 185Hz to 195Hz, the second frequency threshold is 198Hz to 202Hz, and the third frequency threshold is 205Hz to 215Hz.

8. A concrete vibration construction control system based on AR visual laser scanning according to claim 7, characterized in that: The process of generating the deep correction guide line is as follows: the real-time spatial coordinates of the vibrator are used as the starting point of the guide line, and the coordinates of the standard design vibration point in the three-dimensional reference model are used as the ending point of the guide line. The two points are connected to form the guide line and superimposed on the AR real scene.

9. A concrete vibration construction control system based on AR visual laser scanning according to claim 1, characterized in that: The vibratory rod sensor group consists of a depth sensor, a vibration frequency sensor, and a timing module; the 3D laser scanner, AR vision device, and main controller use 5G and WiFi dual-mode communication to transmit data.

Images