A method and system for monitoring a precast component production line

Abstract

The application discloses a prefabricated component production line monitoring method and system, and belongs to the technical field of industrial monitoring. The method comprises the following steps: acquiring a three-dimensional point cloud data set of a prefabricated component on a production line, and determining the spatial position coordinates of the component according to the three-dimensional point cloud data set; comparing the spatial position coordinates of the component with pre-established standard position information to obtain offset vector data; if the offset vector data exceeds a threshold range, a deviation correction control signal is generated; after the component is adjusted, the corrected spatial position information of the component is acquired; then the theoretical position coordinates of steel bar binding are calculated; the theoretical position coordinates of steel bar binding are compared with the actual installation position data of the steel bar binding to determine a matching accuracy value; if the matching accuracy value is lower than a preset accuracy threshold, optimization coordinate information is generated. The prefabricated component production line monitoring method and system solve the problems of low production efficiency and low product quality in the current prefabricated component production mode.

Description

Technical Field

[0001] This invention relates to the field of industrial monitoring technology, and in particular to a method and system for monitoring a prefabricated component production line. Background Technology

[0002] Precast component production, as a core area of ​​building industrialization, plays a crucial role in improving building quality and efficiency. This field is directly related to the stability and safety of building structures, and is also a key link in promoting the modernization and intelligent transformation of the construction industry. With the continuous growth of industrialization demands, the monitoring of precast component production lines has become a key direction for industry development, and its importance is self-evident.

[0003] However, current production line monitoring methods have revealed numerous shortcomings in practical applications. Traditional methods often rely on manual operation and experience-based judgment, which is not only time-consuming and labor-intensive but also easily affected by human factors, making it difficult to detect and correct deviations in the production process in a timely manner. A deeper problem lies in the lack of systematic control over the entire production process, particularly in data acquisition and real-time analysis. This results in poor coordination between production stages, making it difficult to meet the demands of large-scale, high-precision production. This issue further affects subsequent processes, such as the accuracy of rebar tying, because positional deviations directly interfere with the spatial matching of these processes, causing quality defects and rework risks. Summary of the Invention

[0004] To overcome the shortcomings of existing technologies, this invention provides a method and system for monitoring precast component production lines to solve the aforementioned problems.

[0005] The technical solution adopted by this invention to solve its technical problem is: a method for monitoring a precast component production line, comprising the following steps:

[0006] S1: Obtain the 3D point cloud dataset of the prefabricated components on the production line, and determine the spatial coordinates of the components based on the 3D point cloud dataset;

[0007] S2: Compare the spatial coordinates of the component with the pre-established standard position information, calculate the actual position offset of the component, and obtain the offset vector data;

[0008] S3: If the offset vector data exceeds the threshold range, a deviation correction control signal is generated and transmitted to the automated monitoring platform;

[0009] S4: Receive the deviation correction control signal through the automated monitoring platform, drive the robotic arm to perform position adjustment operation, and obtain the corrected spatial position information of the component after the component is adjusted;

[0010] S5: Based on the corrected spatial position information of the component, calculate the theoretical position coordinates of the reinforcing bar binding;

[0011] S6: The theoretical position coordinates of the rebar binding are compared with the actual installation position data of the rebar binding in the precast components on the production line to determine the matching accuracy value between the theoretical position coordinates and the actual installation position. If the matching accuracy value is lower than the preset accuracy threshold, optimized coordinate information is generated and transmitted to the intelligent management platform.

[0012] Preferably, in step S1, the three-dimensional point cloud dataset is acquired using a high-precision sensor array and a laser scanning device;

[0013] After acquisition, the voxel filtering algorithm is used to denoise the three-dimensional point cloud dataset to remove outliers and noise, resulting in a denoised point cloud dataset.

[0014] The denoised point cloud dataset is filtered using a statistical filtering algorithm to smooth the point cloud data and obtain an optimized point cloud dataset.

[0015] Based on the optimized point cloud dataset, a pre-established homogeneous transformation matrix is ​​applied to transform the optimized point cloud data from the local coordinate system to the global coordinate system, resulting in point cloud data in the global coordinate system.

[0016] The spatial coordinates of the component are determined based on the point cloud data in the global coordinate system.

[0017] Optionally, in step S2, the step of comparing the spatial coordinates of the component with pre-established standard position information includes:

[0018] A preliminary spatial location dataset of prefabricated components is generated using their spatial coordinates.

[0019] The iterative nearest point algorithm is used to register the preliminary spatial location dataset with the reference point cloud in the pre-established standard location database, resulting in a registration error matrix composed of translation components of multiple registration points. The translation components of the registration points are the differences between the coordinates in the preliminary spatial location dataset and the coordinates of the reference points they are registered with.

[0020] Specifically, in step S2, the step of calculating the actual position offset of the component and obtaining the offset vector data includes:

[0021] The translation component of each registration point is extracted from the registration error matrix. The translation component represents the actual position offset of the prefabricated component in the global coordinate system.

[0022] For the translation component, an XYZ tri-axis orthogonal decomposition is performed in the global coordinate system to obtain offset vector data containing X-axis, Y-axis and Z-axis components.

[0023] It is worth noting that in step S3, the X-axis component and Y-axis component of each registration point are compared with a pre-established threshold range point by point; if the X-axis component exceeds the first threshold range, the X-axis deviation is recorded; if the Y-axis component exceeds the second threshold range, the Y-axis deviation is recorded.

[0024] Based on the deviation identifier, the coordinate points with the deviation identifier are used as input features, and a pre-trained clustering analysis module is used to perform three-class clustering to obtain the clustering results; wherein the clustering results include X-axis deviation, Y-axis deviation and composite deviation;

[0025] Based on the clustering results and the magnitude of the deviation, a corresponding deviation correction control signal is generated.

[0026] Preferably, in step S4, a control channel is selected based on the clustering results. The clustering results of the X-axis deviation correspond to the first axis control channel, the clustering results of the Y-axis deviation correspond to the second axis control channel, and the clustering results of the composite deviation correspond to the dual-axis control channel; the displacement of the robotic arm is the magnitude of the deviation.

[0027] Optionally, in step S5, the pre-established complete component design coordinates are extracted, and the least squares method is used to calculate the transformation matrix between the corrected component spatial position information and the design coordinates;

[0028] The design coordinates are transformed according to the transformation matrix to obtain first coordinate data aligned with the corrected component spatial position information, wherein the first coordinate data contains complete component information.

[0029] Based on the spatial position values ​​of the rebar tying points in the first coordinate data, a target position table for the rebar tying points is generated, which stores the theoretical position coordinates of the rebar tying points.

[0030] Specifically, in step S6, a high-precision laser scanner is used to acquire three-dimensional point cloud data of the actual installation position of the steel reinforcement binding in the prefabricated components on the production line, and the data is stored as an initial dataset.

[0031] Noise is removed from the initial dataset to obtain a denoised point cloud dataset, and the denoised point cloud dataset is then converted to actual coordinates in the global coordinate system.

[0032] Obtain the theoretical position coordinates derived from the target position table, and iteratively register the actual coordinates and the theoretical position coordinates using the ICP algorithm accelerated by kd-tree to obtain the transformation matrix and the corresponding residual matrix.

[0033] If the F-norm of the residual matrix exceeds a preset threshold, the coordinate difference in the direction of the maximum deviation in the residual matrix is ​​extracted, and the actual coordinates are adjusted according to the coordinate difference in the direction of the maximum deviation to obtain optimized coordinate information.

[0034] A precast component production line monitoring system is provided to implement the aforementioned precast component production line monitoring method.

[0035] The beneficial effects of this invention are as follows: In the precast component production line monitoring method, by acquiring three-dimensional point cloud data of the component, the spatial position coordinates of the component are determined. Offset vector data is calculated by comparing this data with standard position information. A correction control signal is generated based on the deviation threshold, driving the robotic arm to perform position adjustment operations. Based on the corrected component position information, the theoretical position coordinates of the rebar binding are calculated. These coordinates are compared with the actual installation position data to determine the matching accuracy value. If the accuracy value is lower than the accuracy threshold, optimized coordinate information is generated. This invention achieves automated position correction and rebar binding accuracy control in the precast component production process, improving production efficiency and quality. Attached Figure Description

[0036] Figure 1 This is a flowchart of a prefabricated component production line monitoring method in one embodiment of the present invention;

[0037] Figure 2 This is a flowchart of step S2 in one embodiment of the present invention. Detailed Implementation

[0038] The specific embodiments of the present invention will be further described below with reference to the accompanying drawings. It should be noted that these descriptions are for the purpose of aiding understanding the present invention, but do not constitute a limitation thereof. Furthermore, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0039] like Figure 1 and 2 As shown, a method for monitoring a precast component production line includes the following steps:

[0040] S1: Obtain the 3D point cloud dataset of the prefabricated components on the production line, and determine the spatial coordinates of the components based on the 3D point cloud dataset;

[0041] S2: Compare the spatial coordinates of the component with the pre-established standard position information, calculate the actual position offset of the component, and obtain the offset vector data;

[0042] S3: If the offset vector data exceeds the threshold range, a deviation correction control signal is generated and transmitted to the automated monitoring platform;

[0043] S4: Receive the deviation correction control signal through the automated monitoring platform, drive the robotic arm to perform position adjustment operation, and obtain the corrected spatial position information of the component after the component is adjusted;

[0044] S5: Based on the corrected spatial position information of the component, calculate the theoretical position coordinates of the reinforcing bar binding;

[0045] S6: The theoretical position coordinates of the rebar binding are compared with the actual installation position data of the rebar binding in the precast components on the production line to determine the matching accuracy value between the theoretical position coordinates and the actual installation position. If the matching accuracy value is lower than the preset accuracy threshold, optimized coordinate information is generated and transmitted to the intelligent management platform.

[0046] In the precast component production line monitoring method, the spatial coordinates of the component are determined by acquiring its three-dimensional point cloud data. These coordinates are then compared with standard position information to calculate the offset vector data. A correction control signal is generated based on the deviation threshold, driving a robotic arm to perform position adjustment. Based on the corrected component position information, the theoretical coordinates of the rebar binding are calculated and compared with actual installation position data to determine the matching accuracy value. If the accuracy is below the threshold, optimized coordinate information is generated. This invention achieves automated position correction and rebar binding accuracy control during precast component production, improving production efficiency and quality.

[0047] It is worth noting that in step S1, the three-dimensional point cloud dataset is acquired through a high-precision sensor array and laser scanning equipment;

[0048] After acquisition, the voxel filtering algorithm is used to denoise the three-dimensional point cloud dataset to remove outliers and noise, resulting in a denoised point cloud dataset.

[0049] The denoised point cloud dataset is filtered using a statistical filtering algorithm to smooth the point cloud data and obtain an optimized point cloud dataset.

[0050] Based on the optimized point cloud dataset, a pre-established homogeneous transformation matrix is ​​applied to transform the optimized point cloud data from the local coordinate system to the global coordinate system, resulting in point cloud data in the global coordinate system.

[0051] The spatial coordinates of the component are determined based on the point cloud data in the global coordinate system.

[0052] Upon acquisition, a preliminary quality assessment is performed on the 3D point cloud dataset to determine its completeness and accuracy, yielding an assessment result. This preliminary quality assessment checks the completeness and accuracy of the point cloud. Completeness is assessed by examining the point cloud density distribution map to determine if there are any obvious data gaps; accuracy is verified by comparing the point cloud with the geometric parameters of the design drawings to confirm that the point cloud does not deviate from the actual dimensions by more than 2 millimeters. The assessment results generate a report indicating missing areas or parts with excessive errors, facilitating subsequent supplementary data collection or correction.

[0053] In this embodiment, a voxel filtering algorithm is used for denoising. Assume the original point cloud contains noisy points, such as outliers caused by environmental reflections. Voxel filtering divides the point cloud into a voxel grid with sides of 5 mm. The average value within each voxel is taken, and voxels with fewer than 3 points are removed, thus eliminating isolated noise points. The denoised point cloud data size is reduced by approximately 20%, while retaining key geometric features of the beam surface, improving data processing efficiency.

[0054] In one possible implementation, statistical filtering algorithms further smooth the point cloud data. For the denoised point cloud dataset, statistical filtering calculates the average neighborhood distance for each point and removes outliers that deviate from the standard deviation by more than twice. For example, the surface of a beam may have slight undulations due to scanning jitter; statistical filtering smooths these fluctuations, making the point cloud surface more continuous and facilitating subsequent geometric analysis.

[0055] Point cloud coordinate system transformation relies on a pre-established homogeneous transformation matrix. Assuming the beam point cloud is initially in the local coordinate system of the laser scanner, the transformation matrix is ​​calculated using known global coordinate system reference points, such as control points measured on-site. The local point cloud coordinates are then multiplied by this matrix to transform to the global coordinate system.

[0056] In one possible implementation, when calculating the spatial coordinates of a prefabricated component, key feature points of the component, such as the center point or the location of connecting holes, can be extracted from the point cloud data in the global coordinate system. Based on the point cloud in the global coordinate system, the three-dimensional coordinates of these feature points are calculated; for example, the coordinates of the component's center point are (10.5, 2.3, 1.8) meters.

[0057] Preferably, in step S2, the step of comparing the spatial coordinates of the component with pre-established standard position information includes:

[0058] A preliminary spatial location dataset of prefabricated components is generated using their spatial coordinates.

[0059] The iterative nearest point algorithm is used to register the preliminary spatial location dataset with the reference point cloud in the pre-established standard location database, resulting in a registration error matrix composed of translation components of multiple registration points. The translation components of the registration points are the differences between the coordinates in the preliminary spatial location dataset and the coordinates of the reference points they are registered with.

[0060] The preliminary spatial location dataset includes multiple spatial coordinates of the prefabricated component. These coordinates, when combined, represent the position of the prefabricated component in the global coordinate system, providing a basis for subsequent registration. In one possible implementation, when performing point cloud registration using an iterative nearest-point algorithm, the preliminary spatial location dataset is compared with a reference point cloud in a standard location database. The iterative nearest-point algorithm iteratively optimizes the transformation matrix by finding the closest point pairs in the preliminary spatial location dataset and the reference point cloud until the two spatial location datasets and the reference point cloud coincide. After registration is complete, a registration error matrix is ​​generated, recording the distance deviation between each pair of points.

[0061] Specifically, the reference point cloud is typically generated based on design drawings. In this embodiment, a coordinate change table of prefabricated components is pre-stored during the entire production line process under normal conditions. This change table arranges the standard coordinate groups of the prefabricated components according to time sequence, thereby obtaining the coordinate changes of the prefabricated components on the production line throughout the entire process. Each set of standard coordinate groups corresponds to a time node in the production process. When the 3D point cloud dataset of the prefabricated components on the production line is collected in step S1, a time node is also formed. By matching time nodes, the corresponding time node of the production process can be found. Then, based on this time node of the production process, the corresponding standard coordinate group of the prefabricated component can be found, and this set of standard coordinate groups is used as the reference point cloud.

[0062] Optionally, in step S2, the step of calculating the actual position offset of the component to obtain the offset vector data includes:

[0063] The translation component of each registration point is extracted from the registration error matrix. The translation component represents the actual position offset of the prefabricated component in the global coordinate system.

[0064] For the translation component, an XYZ tri-axis orthogonal decomposition is performed in the global coordinate system to obtain offset vector data containing X-axis, Y-axis and Z-axis components.

[0065] Translation components represent the positional deviation of a component in the global coordinate system. These components reflect the difference between the component's actual position during scanning and its designed position. In one possible implementation, the translation components can be decomposed into three independent components when performing an XYZ tri-axis orthogonal decomposition. For example, the X-axis component might be 2.5 mm, the Y-axis component 1.8 mm, and the Z-axis component 0.9 mm. This decomposition facilitates precise correction along each axis when adjusting the position of the equipment or component. The decomposition process relies on a right-handed Cartesian coordinate system to ensure directional consistency.

[0066] Specifically, in step S3, for each registration point, the X-axis component and Y-axis component are compared with a pre-established threshold range point by point; if the X-axis component exceeds the first threshold range, the X-axis deviation indicator is recorded; if the Y-axis component exceeds the second threshold range, the Y-axis deviation indicator is recorded.

[0067] Based on the deviation identifier, the coordinate points with the deviation identifier are used as input features, and a pre-trained clustering analysis module is used to perform three-class clustering to obtain the clustering results; wherein the clustering results include X-axis deviation, Y-axis deviation and composite deviation;

[0068] Based on the clustering results and the magnitude of the deviation, a corresponding deviation correction control signal is generated.

[0069] Suppose that in a specific scenario, the extracted X-axis component of a coordinate point is 3.2 mm and the Y-axis component is 1.5 mm. These values ​​will be used for subsequent threshold comparisons. In this embodiment, since the components are transported on a conveyor belt and will be transported close to the conveyor belt, the Z-axis component is not considered.

[0070] When performing threshold comparisons, a range of three standard deviations based on historical data can be preset, for example, the first threshold range is -2.0 to 2.0 mm, and the second threshold range is -1.8 to 1.8 mm.

[0071] If the X-axis component of a point is 3.2 mm, it clearly exceeds the first threshold range, and an X-axis deviation indicator will be recorded. Similarly, if the Y-axis component also exceeds the second threshold range, a Y-axis deviation indicator will be recorded. This indicator method helps to quickly locate the specific direction and type of deviation, providing a clear basis for subsequent processing.

[0072] When performing cluster analysis based on deviation labels, coordinate points with deviation labels can be used as input, and a pre-trained cluster analysis module can be used for classification. Through cluster analysis, these points are divided into three categories, corresponding to X-axis deviation, Y-axis deviation, and combined deviation. The X-axis deviation category indicates that the component is offset only in the X-axis direction, the Y-axis deviation category indicates that the component is offset only in the Y-axis direction, and the combined deviation category indicates that the component is offset in both the X-axis and Y-axis directions. This classification method can categorize different types of deviation problems, facilitating targeted processing.

[0073] When generating deviation correction control signals, specific correction strategies can be formulated based on clustering results and deviation magnitudes. Assuming the average X-axis deviation is 2.5 mm and the average Y-axis deviation is 1.7 mm, the system will generate corresponding control signals based on the clustering results of the composite deviation, the X-axis deviation magnitude of 2.5 mm, and the Y-axis deviation magnitude of 1.7 mm, guiding the equipment to adjust 2.5 mm in the reverse direction along the X-axis and 1.7 mm in the reverse direction along the Y-axis.

[0074] It is worth noting that in step S4, the control channel is selected based on the clustering results. The clustering results of the X-axis deviation correspond to the first axis control channel, the clustering results of the Y-axis deviation correspond to the second axis control channel, and the clustering results of the composite deviation correspond to the dual-axis control channel; the displacement of the robotic arm is the magnitude of the deviation.

[0075] Preferably, in step S5, the pre-established complete component design coordinates are extracted, and the least squares method is used to calculate the transformation matrix between the corrected component spatial position information and the design coordinates;

[0076] The design coordinates are transformed according to the transformation matrix to obtain first coordinate data aligned with the corrected component spatial position information, wherein the first coordinate data contains complete component information.

[0077] Based on the spatial position values ​​of the rebar tying points in the first coordinate data, a target position table for the rebar tying points is generated, which stores the theoretical position coordinates of the rebar tying points.

[0078] The design coordinate system is usually a virtual reference frame based on the component's theoretical model, containing complete positional information of the component under ideal conditions.

[0079] In one possible implementation, assuming a steel structure component fabrication scenario, the design coordinates include the theoretical position coordinates of multiple key nodes on the component. For example, the coordinates of the four endpoints of a component might be A(0,0,0), B(1000,0,0), C(1000,2000,0), and D(0,2000,0), in millimeters. The complete component design coordinates are used to compare with the actual, corrected component position. This approach ensures that the subsequent calculated transformation matrix is ​​based on the theoretical design, exhibiting a high degree of reference consistency.

[0080] The process of calculating the transformation matrix using the least squares method can be understood as an algorithm for optimizing spatial alignment. The core of the least squares method lies in finding the optimal transformation relationship between the actual component positions and the design coordinates through mathematical methods. In one embodiment, assuming the corrected component position data are A'(10,5,2), B'(1012,3,1), and C'(1015,2008,3), the system uses the deviations between these actual points and the design coordinate points to calculate a transformation matrix using the least squares method, which includes translation, rotation, and scaling parameters. This matrix can bring the actual position data as close as possible to the design coordinates, reducing overall error.

[0081] The step of transforming the design coordinates using a transformation matrix to obtain the first coordinate data is primarily to convert the design coordinates into coordinates aligned with the actual position. This process ensures that the design coordinates accurately reflect the position of the actual component, providing an accurate reference for subsequent processes.

[0082] The generated first coordinate data contains complete component information. This means that during the transformation process, the coordinates of the rebar tying in the design coordinates are also transformed into coordinates aligned with the corrected spatial position information of the component. In this scheme, if a component hasn't reached the final process flow during production, its structure on the production line is incomplete, indicating that all processes haven't been completed. Therefore, the corrected spatial position information also represents an incomplete component structure. Thus, pre-established complete component design coordinates are needed. After coordinate transformation, the design coordinates are aligned with the corrected spatial position information of the component. Then, the theoretical positions of the rebar tying in the upcoming process are found from the first coordinate data. The system stores this first coordinate data as a temporary dataset for easy retrieval in subsequent steps. When generating the target position table of rebar tying points, the system extracts the specific position coordinates of the rebar tying points based on the first coordinate data and organizes them into a table format. In one embodiment, assuming there are 8 rebar tying points on the component, the system will obtain the theoretical positions of these points based on the first coordinate data, such as points P1(200,500,0), P2(200,1500,0), etc., and generate a table containing coordinate values ​​and numbers.

[0083] Optionally, in step S6, three-dimensional point cloud data of the actual installation position of the steel reinforcement binding in the prefabricated components on the production line is acquired by a high-precision laser scanner and stored as an initial dataset.

[0084] Noise is removed from the initial dataset to obtain a denoised point cloud dataset, and the denoised point cloud dataset is then converted to actual coordinates in the global coordinate system.

[0085] Obtain the theoretical position coordinates derived from the target position table, and iteratively register the actual coordinates and the theoretical position coordinates using the ICP algorithm accelerated by kd-tree to obtain the transformation matrix and the corresponding residual matrix.

[0086] If the F-norm of the residual matrix exceeds a preset threshold, the coordinate difference in the direction of the maximum deviation in the residual matrix is ​​extracted, and the actual coordinates are adjusted according to the coordinate difference in the direction of the maximum deviation to obtain optimized coordinate information.

[0087] It should be noted that the initial dataset often introduces noise due to ambient light, surface reflection, or equipment vibration, affecting the accuracy of subsequent analysis. Therefore, the system employs statistical filtering to remove noise. By calculating the neighborhood density of each point and eliminating isolated points that deviate too far from the average distance, a denoised point cloud dataset is obtained, potentially reducing the number of points from 10 million to 9 million, significantly improving data accuracy. Specifically, the denoised point cloud dataset needs to be converted to actual coordinates in the global coordinate system. This conversion ensures that all point cloud data is consistent with the global coordinate system, facilitating subsequent registration.

[0088] After obtaining the theoretical position coordinates from the target position table, the system iteratively registers the actual coordinates with the theoretical position coordinates using the ICP algorithm accelerated by kd-trees. The kd-tree optimizes the point pair matching efficiency through spatial partitioning. For example, when processing 100,000 points, the registration time is reduced from several minutes to several seconds. The registration result generates a transformation matrix that describes the translation and rotation relationship between the actual coordinates and the theoretical position coordinates. At the same time, it outputs a residual matrix that reflects the deviation of each point between the actual coordinates and the theoretical position coordinates.

[0089] If the F-norm of the residual matrix exceeds a threshold, such as 0.5 mm, it indicates that the registration error is too large. In this case, the system extracts the coordinate difference along the direction of maximum deviation. For example, if the Z-axis deviation reaches 0.8 mm, the original coordinates of a rebar binding point are (1000.5, 2000.3, 100.8) mm. Based on the coordinate difference along the direction of maximum deviation, these coordinates are adjusted to (1000.5, 2000.3, 100.0) mm, resulting in optimized coordinate information and reducing the error to within the threshold. This optimized coordinate information can be used to guide the robotic arm to bind rebars at precise positions on the component.

[0090] A precast component production line monitoring system is provided to implement the aforementioned precast component production line monitoring method.

[0091] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to the described embodiments. For those skilled in the art, various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention, and these variations still fall within the protection scope of the present invention.

Claims

1. A method for monitoring a precast component production line, characterized in that, Includes the following steps: S1: Obtain the 3D point cloud dataset of the prefabricated components on the production line, and determine the spatial coordinates of the components based on the 3D point cloud dataset; S2: Compare the spatial coordinates of the component with the pre-established standard position information, calculate the actual position offset of the component, and obtain the offset vector data; S3: If the offset vector data exceeds the threshold range, a deviation correction control signal is generated and transmitted to the automated monitoring platform; S4: Receive the deviation correction control signal through the automated monitoring platform, drive the robotic arm to perform position adjustment operation, and obtain the corrected spatial position information of the component after the component is adjusted; S5: Based on the corrected spatial position information of the component, calculate the theoretical position coordinates of the reinforcing bar binding; S6: The theoretical position coordinates of the rebar binding are compared with the actual installation position data of the rebar binding in the precast components on the production line to determine the matching accuracy value between the theoretical position coordinates and the actual installation position. If the matching accuracy value is lower than the preset accuracy threshold, optimized coordinate information is generated and transmitted to the intelligent management platform.

2. The method for monitoring a precast component production line according to claim 1, characterized in that: In step S1, the three-dimensional point cloud dataset is acquired using a high-precision sensor array and a laser scanning device; After acquisition, the voxel filtering algorithm is used to denoise the three-dimensional point cloud dataset to remove outliers and noise, resulting in a denoised point cloud dataset. The denoised point cloud dataset is filtered using a statistical filtering algorithm to smooth the point cloud data and obtain an optimized point cloud dataset. Based on the optimized point cloud dataset, a pre-established homogeneous transformation matrix is ​​applied to transform the optimized point cloud data from the local coordinate system to the global coordinate system, resulting in point cloud data in the global coordinate system. The spatial coordinates of the component are determined based on the point cloud data in the global coordinate system.

3. The method for monitoring a precast component production line according to claim 1, characterized in that: In step S2, the step of comparing the spatial coordinates of the component with pre-established standard position information includes: A preliminary spatial location dataset of prefabricated components is generated using their spatial coordinates. The iterative nearest point algorithm is used to register the preliminary spatial location dataset with the reference point cloud in the pre-established standard location database, resulting in a registration error matrix composed of translation components of multiple registration points. The translation components of the registration points are the differences between the coordinates in the preliminary spatial location dataset and the coordinates of the reference points they are registered with.

4. The method for monitoring a precast component production line according to claim 3, characterized in that: In step S2, the step of calculating the actual position offset of the component and obtaining the offset vector data includes: The translation component of each registration point is extracted from the registration error matrix. The translation component represents the actual position offset of the prefabricated component in the global coordinate system. For the translation component, an XYZ tri-axis orthogonal decomposition is performed in the global coordinate system to obtain offset vector data containing X-axis, Y-axis and Z-axis components.

5. The method for monitoring a precast component production line according to claim 4, characterized in that: In step S3, for each registration point, the X-axis component and Y-axis component are compared with a pre-established threshold range point by point; if the X-axis component exceeds the first threshold range, the X-axis deviation indicator is recorded; if the Y-axis component exceeds the second threshold range, the Y-axis deviation indicator is recorded. Based on the deviation identifier, the coordinate points with the deviation identifier are used as input features, and a pre-trained clustering analysis module is used to perform three-class clustering to obtain the clustering results. The clustering results include X-axis bias, Y-axis bias, and composite bias. Based on the clustering results and the magnitude of the deviation, a corresponding deviation correction control signal is generated.

6. The method for monitoring a precast component production line according to claim 5, characterized in that: In step S4, a control channel is selected based on the clustering results. The clustering results of the X-axis deviation correspond to the first axis control channel, the clustering results of the Y-axis deviation correspond to the second axis control channel, and the clustering results of the composite deviation correspond to the dual-axis control channel. The displacement of the robotic arm is the magnitude of the deviation.

7. The method for monitoring a precast component production line according to claim 1, characterized in that: In step S5, the pre-established complete component design coordinates are extracted, and the least squares method is used to calculate the transformation matrix between the corrected component spatial position information and the design coordinates. The design coordinates are transformed according to the transformation matrix to obtain first coordinate data aligned with the corrected component spatial position information, wherein the first coordinate data contains complete component information. Based on the spatial position values ​​of the rebar tying points in the first coordinate data, a target position table for the rebar tying points is generated, which stores the theoretical position coordinates of the rebar tying points.

8. The method for monitoring a precast component production line according to claim 1, characterized in that: In step S6, a high-precision laser scanner is used to acquire three-dimensional point cloud data of the actual installation position of the steel reinforcement binding in the prefabricated components on the production line, and the data is stored as an initial dataset. Noise is removed from the initial dataset to obtain a denoised point cloud dataset, and the denoised point cloud dataset is then converted to actual coordinates in the global coordinate system. Obtain the theoretical position coordinates derived from the target position table, and iteratively register the actual coordinates and the theoretical position coordinates using the ICP algorithm accelerated by kd-tree to obtain the transformation matrix and the corresponding residual matrix. If the F-norm of the residual matrix exceeds a preset threshold, the coordinate difference in the direction of the maximum deviation in the residual matrix is ​​extracted, and the actual coordinates are adjusted according to the coordinate difference in the direction of the maximum deviation to obtain optimized coordinate information.

9. A monitoring system for a precast component production line, characterized in that: Used to implement the precast component production line monitoring method according to any one of claims 1-8.

Images