An intelligent detection method and assembly system for the opening degree of segment capping blocks

Abstract

The present invention discloses an intelligent detection method and assembly system for the aperture of a segment capping block, belonging to the technical field of shield tunnel construction control. The method comprises: matching the point cloud data of the left adjacent block acquired by the left visual sensor with the target point cloud data of the left adjacent block to obtain a posture transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system; multiplying the posture transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system with the posture transformation matrix of the left visual sensor coordinate system relative to the global coordinate system to obtain a transformation matrix from the left local coordinate system to the global coordinate system; collecting the coordinates of the vertices on the contact surface between the left adjacent block and the segment capping block in the left local coordinate system, multiplying them by the transformation matrix to obtain the global coordinates of the left vertices, and calculating the actual aperture of the segment capping block using the global coordinates of the left and right vertices. The present invention has a low detection error and effectively improves construction efficiency and safety.

Description

Technical Field

[0001] The present invention belongs to the technical field of shield tunnel construction control, and more specifically, relates to an intelligent detection method and assembly system for the opening degree of a segment capping block. Background Art

[0002] Shield tunneling, the mainstream construction technology for modern tunneling, relies on segment assembly as its core process. The quality of segment assembly is directly related to the overall safety of the tunnel structure. The capping block, a key component in completing the segment ring, requires special process sensitivity for assembly precision. Engineering practice has shown that when the capping block opening (i.e., the distance between the corner points of the inner arc surface of two adjacent blocks) does not meet the design and construction experience threshold, forced assembly will lead to abnormal stress concentrations on the segment contact surface, causing major quality defects such as structural cracking and waterstop tearing, seriously threatening the waterproof performance and structural durability of the tunnel project.

[0003] Existing technologies for detecting the opening of capping blocks still rely on traditional manual operation, which presents significant technical limitations: First, measurement accuracy is limited by subjective judgment, which can easily lead to measurement errors under complex working conditions. Second, the limited space in shield tunnels restricts the number and location of sensors, making it difficult to fully cover the capping opening. Third, for large-diameter and ultra-large-diameter shield tunneling projects, workers must repeatedly climb to high-risk working platforms more than 10 meters above the ground to observe three-dimensional spatial relationships, severely restricting construction efficiency. Fourth, the high-risk high-altitude working environment exposes construction workers to multiple safety hazards, including but not limited to falls from height and human-machine collisions. As urban underground space development in my country develops towards deeper layers and larger cross-sections, traditional manual inspection methods are no longer able to meet the dual demands of construction accuracy and timeliness for new shield tunneling projects.

[0004] It can be seen that the existing capping block opening detection technology has technical problems such as large error, low efficiency and poor safety. Summary of the Invention

[0005] In response to the above defects or improvement needs of the prior art, the present invention provides an intelligent detection method and assembly system for the opening degree of pipe segment capping blocks, thereby solving the technical problems of large errors, low efficiency and poor safety in the existing capping block opening degree detection technology.

[0006] To achieve the above-mentioned object, according to one aspect of the present invention, a method for intelligently detecting the opening degree of a segment capping block is provided, comprising:

[0007] A left local coordinate system is established with the center of the bolt hole on the inner arc surface of the left adjacent block of the segment capping block as the circle center, a right local coordinate system is established with the center of the bolt hole on the inner arc surface of the right adjacent block of the segment capping block as the circle center, and a global coordinate system is established with the rotation center of the segment assembly machine as the circle center;

[0008] Match the left adjacent block point cloud data acquired by the left visual sensor with the left adjacent block target point cloud data to obtain the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system; match the right adjacent block point cloud data acquired by the right visual sensor with the right adjacent block target point cloud data to obtain the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system;

[0009] Multiply the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system with the pose transformation matrix of the left visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the left local coordinate system to the global coordinate system. Multiply the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system with the pose transformation matrix of the right visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the right local coordinate system to the global coordinate system.

[0010] Collect the coordinates of the vertex in the contact surface between the left adjacent block and the segment capping block in the left local coordinate system, multiply it with the transformation matrix from the left local coordinate system to the global coordinate system to obtain the global coordinates of the left vertex; collect the coordinates of the vertex in the contact surface between the right adjacent block and the segment capping block in the right local coordinate system, multiply it with the transformation matrix from the right local coordinate system to the global coordinate system to obtain the global coordinates of the right vertex; use the global coordinates of the left vertex and the global coordinates of the right vertex to calculate the distance between the left vertex and the right vertex to obtain the actual opening degree of the segment capping block.

[0011] Furthermore, the transformation matrix from the left local coordinate system to the global coordinate system is calculated in the same manner as the transformation matrix from the right local coordinate system to the global coordinate system. The transformation matrix from the left local coordinate system to the global coordinate system is calculated as follows:

[0012] Perform nearest neighbor matching on the geometric features of the source point cloud in the left adjacent block point cloud data and the target point cloud in the left adjacent block target point cloud data to establish initial feature matching point pairs. Use singular value decomposition to solve the rigid transformation matrix for the initial feature matching point pairs to complete the coarse registration.

[0013] The rigid transformation matrix is ​​used as the initial value, and the pseudo-Huber loss function between the source point cloud and the target point cloud that have completed the coarse registration is used as the optimization target. Through iterative optimization, when the pseudo-Huber loss function value meets the convergence condition or reaches the preset maximum number of iterations, the iteration is terminated, and the transformation matrix from the local coordinate system on the left to the global coordinate system is obtained.

[0014] Furthermore, the geometric features of the source point cloud and the geometric features of the target point cloud are calculated in the same manner. The geometric features of the source point cloud are calculated as follows:

[0015] The frequency histogram of the source point cloud and its neighborhood point cloud is calculated by the angular eigenvalues ​​of the source point cloud and its neighborhood point cloud to obtain the SPFH descriptor of the source point cloud and its neighborhood point cloud. The SPFH descriptors of all the neighborhood point clouds of the source point cloud are weighted averaged and then added to the SPFH descriptor of the source point cloud to obtain the FPFH descriptor of the source point cloud.

[0016] Perform weighted averaging on the weights of all neighboring point clouds of the source point cloud to obtain neighborhood enhancement features;

[0017] The FPFH descriptor and neighborhood enhancement features of the source point cloud constitute the geometric features of the source point cloud.

[0018] Furthermore, the weight of the neighborhood point cloud is:

[0019] ;

[0020] in, The first point in the source point cloud The coordinates of the points, is the first point in the neighborhood point cloud j The coordinates of the points, is the first point in the neighborhood point cloud j The weight of the point, , is the weight attenuation coefficient, is the bolt hole base weight gain, is the base weight for the points other than the bolt hole in the left adjacent block.

[0021] Furthermore, the pseudo-Huber loss function is:

[0022] ;

[0023] in, is a hyperparameter, is the rotation matrix, is the translation vector, The first point in the target point cloud The coordinates of the points, To enhance the features of the neighborhood, A set of source point clouds for coarse registration.

[0024] According to another aspect of the present invention, a system for intelligently detecting the opening of a segment capping block is provided, comprising a processor configured to execute a method for intelligently detecting the opening of a segment capping block.

[0025] According to another aspect of the present invention, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, a method for intelligently detecting the opening degree of a segment capping block is implemented.

[0026] According to another aspect of the present invention, a segment capping block assembly system is provided, comprising: a left visual sensor, a right visual sensor, an aperture calculation module, and a segment assembly machine;

[0027] The left visual sensor is located on the rotary mechanism of the segment assembly machine. The field of view of the left visual sensor includes the area near the left adjacent block, and is used to collect point cloud data of the left adjacent block and transmit it to the opening calculation module;

[0028] The right visual sensor is located on the rotary mechanism of the segment assembly machine. The field of view of the right visual sensor includes the right adjacent block and is used to collect point cloud data of the right adjacent block and transmit it to the opening calculation module;

[0029] The opening calculation module uses an intelligent detection method for the opening of the segment capping block to calculate the actual opening of the segment capping block;

[0030] The rotating mechanism of the segment assembly machine drives the segment assembly machine to rotate around the rotation center to between the left adjacent block and the right adjacent block;

[0031] The segment assembling machine assembles the segment capping block between the left adjacent block and the right adjacent block when the difference between the actual opening degree of the segment capping block and the theoretical value is within the construction error range.

[0032] Furthermore, the system also includes: a data center, a programmable logic controller and a motion controller;

[0033] The data center transmits a control instruction to the programmable logic controller when the difference between the actual opening degree of the segment capping block and the theoretical value is within the construction error range, and transmits a stop instruction to the programmable logic controller when the difference between the actual opening degree of the segment capping block and the theoretical value is not within the construction error range;

[0034] The programmable logic controller transmits the control instruction to the motion controller, and the motion controller controls the segment assembly machine to assemble the segment capping block between the left adjacent block and the right adjacent block;

[0035] The programmable logic controller transmits a stop instruction to the motion controller, and the motion controller controls the segment assembly machine to stop assembly.

[0036] Furthermore, the system further comprises: a visualization module,

[0037] The visualization module is used to visualize the error between the actual opening of the segment capping block obtained by the opening calculation module and the theoretical value on the Web side, assisting construction personnel in making decisions on assembling the segment capping blocks.

[0038] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0039] (1) The present invention uses the measurement data of the left and right visual sensors to match the target data to obtain the posture transformation matrix, thereby avoiding the measurement accuracy being limited by the subjective judgment of manual measurement and ensuring the measurement accuracy even under complex working conditions. The present invention only requires the data of the left adjacent block and the right adjacent block, thus overcoming the problem of limited space in the shield tunnel. The present invention first establishes local and global coordinates, then measures the point cloud data of the left and right adjacent blocks, calculates the local to global transformation matrix, and finally obtains the contact surface vertex coordinates based on the transformation matrix, thereby efficiently and accurately detecting the actual opening of the segment capping block, ensuring that the subsequent capping block splicing construction is efficient and safe. The detection method of the present invention utilizes the measurement data of the sensor and the intelligent calculation process to efficiently and accurately calculate the opening, meeting the dual requirements of the new shield engineering for construction accuracy and operation timeliness.

[0040] (2) The present invention performs a rough registration followed by iterative optimization to obtain a more accurate transformation matrix, thus ensuring the accuracy of subsequent aperture calculations. The present invention performs a weighted average on the weights of all neighborhood point clouds of the source point cloud to obtain neighborhood enhancement features. The FPFH descriptor of the source point cloud and the neighborhood enhancement features constitute the geometric features of the source point cloud, which are enhanced geometric features. Using the enhanced geometric features for feature matching can improve matching accuracy.

[0041] (3) Point cloud registration relies on local geometric features. The classification of a single point cannot fully characterize its surrounding environment. Directly assigning a fixed weight to the corresponding point (such as 1.0 for points inside a bolt hole and 0.5 for other points) will ignore the neighborhood context information. If there is an error in the classification (such as mistakenly marking a background point as a bolt hole), the fixed weight will amplify the error. Based on this, the present invention proposes a weight allocation method that combines classification with distance attenuation. Even if a point is mistakenly marked as a point inside a bolt hole, its weight will still be attenuated if it is far away from the current point. The present invention uses a dynamic weight adjustment-based FPFH algorithm and an improved ICP algorithm to effectively improve the accuracy of feature matching.

[0042] (4) The system of the present invention uses visual sensors to perform intelligent detection of the opening of the pipe segment capping block, reducing the need for manual measurement and improving construction efficiency and assembly accuracy. For the fully automatic pipe segment assembly task, the present invention designs a set of virtual-physical communication lines to realize the linkage between the capping block opening detection results and the assembly machine control system. When it is detected that the capping block opening gap does not meet the construction requirements, the system can automatically stop the task execution of the fully automatic assembly machine to avoid assembly failure or equipment damage due to insufficient gap, thereby improving construction quality, ensuring the safety of equipment and personnel, and promoting the intelligent and automated development of tunnel construction. For manual assembly tasks, the present invention enables construction personnel to monitor the pipe segment assembly status in real time by visualizing the capping block opening detection data, assisting in the decision-making of whether to perform the capping block assembly operation, and improving construction safety and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 This is a flow chart of a method for intelligently detecting the opening degree of a segment capping block provided by an embodiment of the present invention.

[0044] Figure 2 This is a detailed flow chart provided by Example 1 of the present invention.

[0045] FIG3( a ) is a diagram showing the definition of the XYZ axis coordinate system provided by an embodiment of the present invention.

[0046] FIG3( b ) is a diagram showing the definition of the XY axis coordinate system provided by an embodiment of the present invention.

[0047] Figure 4 Schematic diagram of the opening degree of the segment capping block provided in an embodiment of the present invention.

[0048] Figure 5 This is a sensor layout diagram provided by an embodiment of the present invention.

[0049] Figure 6 This is a schematic diagram of the connection of control modules in the automatic assembly machine control system provided by an embodiment of the present invention.

[0050] Figure 7 It is a schematic diagram of a real-time visualization process of the opening degree of a capping block provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0051] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely for the purpose of explaining the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below may be combined with each other as long as they do not conflict with each other.

[0052] like Figure 1As shown, a method for intelligently detecting the opening degree of a segment capping block includes:

[0053] A left local coordinate system is established with the center of the bolt hole on the inner arc surface of the left adjacent block of the segment capping block as the circle center, a right local coordinate system is established with the center of the bolt hole on the inner arc surface of the right adjacent block of the segment capping block as the circle center, and a global coordinate system is established with the rotation center of the segment assembly machine as the circle center;

[0054] Match the left adjacent block point cloud data acquired by the left visual sensor with the left adjacent block target point cloud data to obtain the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system; match the right adjacent block point cloud data acquired by the right visual sensor with the right adjacent block target point cloud data to obtain the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system;

[0055] Multiply the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system with the pose transformation matrix of the left visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the left local coordinate system to the global coordinate system. Multiply the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system with the pose transformation matrix of the right visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the right local coordinate system to the global coordinate system.

[0056] Collect the coordinates of the vertex in the contact surface between the left adjacent block and the segment capping block in the left local coordinate system, multiply it with the transformation matrix from the left local coordinate system to the global coordinate system to obtain the global coordinates of the left vertex; collect the coordinates of the vertex in the contact surface between the right adjacent block and the segment capping block in the right local coordinate system, multiply it with the transformation matrix from the right local coordinate system to the global coordinate system to obtain the global coordinates of the right vertex; use the global coordinates of the left vertex and the global coordinates of the right vertex to calculate the distance between the left vertex and the right vertex to obtain the actual opening degree of the segment capping block.

[0057] Scan the segment with a high-precision scanner to obtain a complete segment point cloud. From the segment point cloud, obtain the point cloud data of the left adjacent block as the target point cloud data for the left adjacent block. From the segment point cloud, obtain the point cloud data of the right adjacent block as the target point cloud data for the right adjacent block. The high-precision scanner used is a Keyence VL-700, Leica RTC360, REVOPOINT MetroX, Leica BLK360, or REVOPOINT Inspire.

[0058] Example 1

[0059] like Figure 2 As shown, a method for intelligently detecting the opening degree of a segment capping block includes:

[0060] (1) Establish a left local coordinate system with the center of the bolt hole on the inner arc surface of the left adjacent block of the segment capping block as the circle center, establish a right local coordinate system with the center of the bolt hole on the inner arc surface of the right adjacent block of the segment capping block as the circle center, and establish a global coordinate system with the rotation center of the segment assembly machine as the circle center;

[0061] Specifically, as shown in Figure 3 (a) and Figure 3 (b), a local coordinate system is established for the bolt holes of the left adjacent block L1 and the right adjacent block L2, and a global coordinate system is established with the rotation center of the assembly machine. Global Coordinate System The center of the circle The rotation center of the segment assembly machine is located on the plane formed by the hinge points at the ends of the left and right translation cylinders and the rotation center point. The axis is parallel to the side of the translation beam guide rail and is directed vertically upward. The axis is perpendicular to the plane formed by the hinge points of the left and right translation cylinders and the center point of rotation, and its direction is opposite to the direction of shield propulsion. The axes are determined by the right-hand rule. Local coordinate system The center of the circle is at the center of the bolt hole on the inner arc surface of the pipe segment. The axis is the normal vector of the inner arc surface of the segment, and its direction points to the inside of the center of the annular segment. The axis is parallel to the arc axis of the segment and its direction is opposite to the tunnel excavation direction. The axes are determined by the right-hand rule.

[0062] (2) Match the left adjacent block point cloud data acquired by the left visual sensor with the left adjacent block target point cloud data to obtain the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system; match the right adjacent block point cloud data acquired by the right visual sensor with the right adjacent block target point cloud data to obtain the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system;

[0063] Multiply the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system with the pose transformation matrix of the left visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the left local coordinate system to the global coordinate system. Multiply the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system with the pose transformation matrix of the right visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the right local coordinate system to the global coordinate system.

[0064] Specifically, the pose transformation matrix of the left visual sensor coordinate system relative to the global coordinate system and the pose transformation matrix of the right visual sensor coordinate system relative to the global coordinate system are calculated in the same way. The pose transformation matrix of the left visual sensor coordinate system relative to the global coordinate system is calculated as follows:

[0065] After establishing the global and local coordinate systems, the visual sensor and assembly machine structure are accurately calibrated by adding calibration objects. This allows the establishment of a homogeneous transformation matrix between the two, providing a solid foundation for subsequent pose calculations. The calibration transformation methods for eye-outside-hand and eye-in-hand are expressed by the following formulas (1) and (2), respectively:

[0066] (1)

[0067] (2)

[0068] In the above formula, taking the eye-outside-hand calibration method as an example, M represents the homogeneous transformation matrix, . Represents the pose transformation matrix of the assembly task base coordinate system relative to the left visual sensor coordinate system, Represents the pose transformation matrix of the calibration sphere coordinate system relative to the left visual sensor coordinate system, Represents the pose transformation matrix of the assembly machine terminal coordinate system relative to the calibration spherical coordinate system, Represents the pose transformation matrix of the assembly task base coordinate system relative to the assembly machine end coordinate system. By utilizing the 7-DOF motion of the assembly machine multiple times, the pose matrix of the calibration ball relative to the visual sensor and the pose matrix of the assembly machine end effector relative to the assembly task base coordinate system are collected. Since the calibration ball is fixed at the end of the robotic arm, for each movement of the assembly machine, The transformation matrix is ​​the same. Ultimately, the pose transformation matrix of the visual sensor relative to the assembly task base coordinate system is solved. For the eye-in-hand calibration method, the goal is to solve the pose transformation matrix of the assembly machine end point relative to the visual sensor. The solution is similar to the eye-out-of-hand calibration method mentioned above. Represents the pose transformation matrix of the assembly task base coordinate system relative to the calibration spherical coordinate system, Represents the pose transformation matrix of the assembly machine terminal coordinate system relative to the assembly task base coordinate system, Represents the pose transformation matrix of the assembly machine's terminal coordinate system relative to the left visual sensor coordinate system. The global coordinate system is the assembly task base coordinate system.

[0069] The frequency histogram of the source point cloud and its neighborhood point cloud is calculated by the angular eigenvalues ​​of the source point cloud and its neighborhood point cloud to obtain the SPFH descriptor of the source point cloud and its neighborhood point cloud. The SPFH descriptors of all the neighborhood point clouds of the source point cloud are weighted averaged and then added to the SPFH descriptor of the source point cloud to obtain the FPFH descriptor of the source point cloud.

[0070] Perform weighted averaging on the weights of all neighboring point clouds of the source point cloud to obtain neighborhood enhancement features;

[0071] The FPFH descriptor and neighborhood enhancement features of the source point cloud constitute the geometric features of the source point cloud.

[0072] The weight of the neighborhood point cloud is formula (3):

[0073] (3)

[0074] in, The first point in the source point cloud The coordinates of the points, is the first point in the neighborhood point cloud j The coordinates of the points, is the first point in the neighborhood point cloud j The weight of the point, , is the weight attenuation coefficient, is the bolt hole base weight gain, is the base weight for the points other than the bolt hole in the left adjacent block.

[0075] Perform nearest neighbor matching on the geometric features of the source point cloud in the left adjacent block point cloud data and the target point cloud in the left adjacent block target point cloud data to establish initial feature matching point pairs. Use singular value decomposition to solve the rigid transformation matrix for the initial feature matching point pairs to complete the coarse registration.

[0076] The rigid transformation matrix is ​​used as the initial value, and the pseudo-Huber loss function (4) between the source point cloud and the target point cloud that have completed the coarse registration is used as the optimization target. Through iterative optimization, when the pseudo-Huber loss function value meets the convergence condition or reaches the preset maximum number of iterations, the iteration is terminated and the transformation matrix from the local coordinate system on the left to the global coordinate system is obtained.

[0077] The pseudo-Huber loss function is:

[0078] (4)

[0079] in, is a hyperparameter, is the rotation matrix, is the translation vector, The first point in the target point cloud The coordinates of the points, Enhance features for the neighborhood = , is the neighborhood point cloud set of the source point cloud, A set of source point clouds for coarse registration.

[0080] The present invention dynamically assigns weights to each corresponding point and assigns higher matching weights to points with strong features, such as bolt holes, to reduce the contribution of low-quality points and improve the accuracy of feature matching.

[0081] (3) Collect the coordinates of the vertex in the contact surface between the left adjacent block and the segment capping block in the left local coordinate system, multiply them by the transformation matrix from the left local coordinate system to the global coordinate system to obtain the global coordinates of the vertex on the left. Collect the coordinates of the vertex in the contact surface between the right adjacent block and the segment capping block in the right local coordinate system, multiply them by the transformation matrix from the right local coordinate system to the global coordinate system to obtain the global coordinates of the vertex on the right. Use the global coordinates of the left vertex and the global coordinates of the right vertex to calculate the distance between the left vertex and the right vertex to obtain the actual opening degree of the segment capping block.

[0082] When the actual opening of the segment capping block D P and theoretical value D T The difference is within the construction error range [ , ], the capping blocks continue to be assembled, otherwise, the assembly stops. and are the minimum construction error and the maximum construction error, respectively.

[0083] Example 2

[0084] A pipe segment capping block assembly system, comprising: a left visual sensor, a right visual sensor, an aperture calculation module, and a pipe segment assembly machine;

[0085] The opening calculation module uses an intelligent detection method for the opening of the segment capping block to calculate the actual opening of the segment capping block;

[0086] like Figure 4 As shown, the four vertices in the contact surface between the left adjacent block and the segment capping block are 1, 3, 5, and 7, and the four vertices in the contact surface between the right adjacent block and the segment capping block are 2, 4, 6, and 8. Specifically, the coordinates of the eight specified points in the local coordinate system and the theoretical opening gap (theoretical value) are obtained by using the measurement function in the SolidWorks software.

[0087] like Figure 5 As shown, the left visual sensor is located on the rotating mechanism, and is used to collect point cloud data of the left adjacent block and transmit it to the opening calculation module; the right visual sensor is located on the rotating mechanism, and is used to collect point cloud data of the right adjacent block and transmit it to the opening calculation module.

[0088] The sensor's field of view in the present invention does not need to completely cover the entire area of ​​the capping block opening gap, but only needs to be able to identify the adjacent segments on the left and right sides. The detection method of the present invention is suitable for situations where the sensor's field of view is limited, and has wider applicability.

[0089] The slewing mechanism causes the segment assembly machine to rotate around the rotation center to between the left adjacent block and the right adjacent block. When the difference between the actual opening of the segment capping block and the theoretical value is within the construction error range, the segment assembly machine assembles the segment capping block between the left adjacent block and the right adjacent block.

[0090] The surface structured light vision sensor installed on the assembly machine's rotating mechanism is used to obtain point clouds of two adjacent adjacent blocks of the capping segment. Since most tunnel segments at this stage are assembled with staggered seams, the position of the segment capping block is uncertain. The rotating mechanism allows the assembly machine to rotate around the rotation center, ensuring the flexibility of the field of view of the surface structured light vision sensor.

[0091] like Figure 6 As shown, the system also includes: a data center, a programmable logic controller and a motion controller;

[0092] The data center transmits a control instruction to the programmable logic controller when the difference between the actual opening degree of the segment capping block and the theoretical value is within the construction error range, and transmits a stop instruction to the programmable logic controller when the difference between the actual opening degree of the segment capping block and the theoretical value is not within the construction error range;

[0093] The programmable logic controller transmits the control instruction to the motion controller, and the motion controller controls the oil cylinder of the segment assembly machine to assemble the segment capping block between the left adjacent block and the right adjacent block;

[0094] The programmable logic controller transmits a stop instruction to the motion controller, and the motion controller controls the segment assembly machine to stop assembly.

[0095] The data center sends control instructions to the programmable logic controller (PLC) via the SLMP protocol. The PLC transmits the instructions to the motion controller via the PROFIBUS-DP protocol.

[0096] The system also includes: visualization module,

[0097] The visualization module is used to visualize the error between the actual opening of the segment capping block obtained by the opening calculation module and the theoretical value on the Web side, assisting construction personnel in making decisions on assembling the segment capping blocks.

[0098] like Figure 7As shown, the segment model was drawn using SolidWorks software and exported as an STL file. However, WebGL cannot directly import STL file models, so the STL file needs to be converted into an FBX file using Blender and imported into the webpage. Four connecting lines are added as sub-objects of the adjacent segment. Every second, the front-end obtains a set of data from the back-end via WebSocket. Based on the received data, a green text annotation showing the data label is added to the midpoint of each line segment. Before each call, previously drawn line segments and labels in the model are cleared to avoid overlap. The text box will continuously update to display the opening degree, assisting construction personnel in making decisions on the assembly of the segment capping blocks.

[0099] The system of the present invention uses visual sensors to perform intelligent detection of the opening of the capping block of the pipe segment, reducing the need for manual measurement and improving construction efficiency and assembly accuracy. For manual assembly tasks, the present invention enables construction personnel to monitor the assembly status of the pipe segment in real time by visualizing the capping block opening detection data, assisting in the decision-making of whether to perform the capping block assembly operation, and improving construction safety and reliability. For fully automatic pipe segment assembly tasks, the present invention designs a set of virtual-physical communication lines to realize the linkage between the capping block opening detection results and the assembly machine control system. When it is detected that the capping block opening gap does not meet the design standard or construction experience threshold, the system can automatically stop the task execution of the fully automatic assembly machine to avoid assembly failure or equipment damage due to insufficient gap, thereby improving construction quality, ensuring the safety of equipment and personnel, and promoting the intelligent and automated development of tunnel construction.

[0100] It will be easily understood by those skilled in the art that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. An intelligent detection method for the opening degree of a segment capping block, characterized in that: include: A left local coordinate system is established with the center of the bolt hole on the inner arc surface of the left adjacent block of the segment capping block as the circle center, a right local coordinate system is established with the center of the bolt hole on the inner arc surface of the right adjacent block of the segment capping block as the circle center, and a global coordinate system is established with the rotation center of the segment assembly machine as the circle center; Match the left adjacent block point cloud data acquired by the left visual sensor with the left adjacent block target point cloud data to obtain the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system; match the right adjacent block point cloud data acquired by the right visual sensor with the right adjacent block target point cloud data to obtain the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system; Multiply the pose transformation matrix of the left local coordinate system relative to the left visual sensor coordinate system with the pose transformation matrix of the left visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the left local coordinate system to the global coordinate system. Multiply the pose transformation matrix of the right local coordinate system relative to the right visual sensor coordinate system with the pose transformation matrix of the right visual sensor coordinate system relative to the global coordinate system to obtain the transformation matrix from the right local coordinate system to the global coordinate system. Collect the coordinates of the vertex in the contact surface between the left adjacent block and the segment capping block in the left local coordinate system, multiply it with the transformation matrix from the left local coordinate system to the global coordinate system to obtain the global coordinates of the left vertex; collect the coordinates of the vertex in the contact surface between the right adjacent block and the segment capping block in the right local coordinate system, multiply it with the transformation matrix from the right local coordinate system to the global coordinate system to obtain the global coordinates of the right vertex; use the global coordinates of the left vertex and the global coordinates of the right vertex to calculate the distance between the left vertex and the right vertex to obtain the actual opening degree of the segment capping block.

2. The intelligent detection method for the opening degree of a segment capping block according to claim 1, characterized in that: The transformation matrix from the left local coordinate system to the global coordinate system is calculated in the same way as the transformation matrix from the right local coordinate system to the global coordinate system. The transformation matrix from the left local coordinate system to the global coordinate system is calculated as follows: Perform nearest neighbor matching on the geometric features of the source point cloud in the left adjacent block point cloud data and the target point cloud in the left adjacent block target point cloud data to establish initial feature matching point pairs. Use singular value decomposition to solve the rigid transformation matrix for the initial feature matching point pairs to complete the coarse registration. The rigid transformation matrix is ​​used as the initial value, and the pseudo-Huber loss function between the source point cloud and the target point cloud that have completed the coarse registration is used as the optimization target. Through iterative optimization, when the pseudo-Huber loss function value meets the convergence condition or reaches the preset maximum number of iterations, the iteration is terminated, and the transformation matrix from the local coordinate system on the left to the global coordinate system is obtained.

3. The intelligent detection method for the opening degree of a segment capping block according to claim 2, characterized in that: The geometric features of the source point cloud and the target point cloud are calculated in the same way. The geometric features of the source point cloud are calculated as follows: The frequency histogram of the source point cloud and its neighborhood point cloud is calculated by the angular eigenvalues ​​of the source point cloud and its neighborhood point cloud to obtain the SPFH descriptor of the source point cloud and its neighborhood point cloud. The SPFH descriptors of all the neighborhood point clouds of the source point cloud are weighted averaged and then added to the SPFH descriptor of the source point cloud to obtain the FPFH descriptor of the source point cloud. Perform weighted averaging on the weights of all neighboring point clouds of the source point cloud to obtain neighborhood enhancement features; The FPFH descriptor and neighborhood enhancement features of the source point cloud constitute the geometric features of the source point cloud.

4. The intelligent detection method for the opening degree of a segment capping block according to claim 3, characterized in that: The weight of the neighborhood point cloud is: ； in, The first point in the source point cloud The coordinates of the points, is the first point in the neighborhood point cloud j The coordinates of the points, is the first point in the neighborhood point cloud j The weight of the point, , is the weight attenuation coefficient, is the bolt hole base weight gain, is the base weight for the points other than the bolt hole in the left adjacent block.

5. The intelligent detection method for the opening degree of a segment capping block according to claim 4, characterized in that: The pseudo-Huber loss function is: ； in, is a hyperparameter, is the rotation matrix, is the translation vector, The first point in the target point cloud The coordinates of the points, To enhance the features of the neighborhood, A set of source point clouds for coarse registration.

6. An intelligent detection system for the opening degree of a segment capping block, characterized in that: The method comprises a processor configured to execute an intelligent detection method for the opening degree of a segment capping block according to any one of claims 1 to 5.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, an intelligent detection method for the opening degree of a segment capping block according to any one of claims 1 to 5 is implemented.

8. A segment capping block assembly system, characterized in that: include: Left-side visual sensor, right-side visual sensor, aperture calculation module and segment assembly machine; The left visual sensor is located on the rotary mechanism of the segment assembly machine. The field of view of the left visual sensor includes the area near the left adjacent block, and is used to collect point cloud data of the left adjacent block and transmit it to the opening calculation module; The right visual sensor is located on the rotary mechanism of the segment assembly machine. The field of view of the right visual sensor includes the right adjacent block and is used to collect point cloud data of the right adjacent block and transmit it to the opening calculation module; The opening calculation module uses the intelligent detection method for the opening of the segment capping block according to any one of claims 1 to 5 to calculate the actual opening of the segment capping block; The rotating mechanism of the segment assembly machine drives the segment assembly machine to rotate around the rotation center to between the left adjacent block and the right adjacent block; The segment assembling machine assembles the segment capping block between the left adjacent block and the right adjacent block when the difference between the actual opening degree of the segment capping block and the theoretical value is within the construction error range.

9. The segment capping block assembly system according to claim 8, characterized in that: The system also includes: a data center, a programmable logic controller and a motion controller; The data center transmits a control instruction to the programmable logic controller when the difference between the actual opening degree of the segment capping block and the theoretical value is within the construction error range, and transmits a stop instruction to the programmable logic controller when the difference between the actual opening degree of the segment capping block and the theoretical value is not within the construction error range; The programmable logic controller transmits the control instruction to the motion controller, and the motion controller controls the segment assembly machine to assemble the segment capping block between the left adjacent block and the right adjacent block; The programmable logic controller transmits a stop instruction to the motion controller, and the motion controller controls the segment assembly machine to stop assembly.

10. The segment capping block assembly system according to claim 8, characterized in that: The system further comprises: a visualization module, The visualization module is used to visualize the error between the actual opening of the segment capping block obtained by the opening calculation module and the theoretical value on the Web side, assisting construction personnel in making decisions on assembling the segment capping blocks.

Images