Automatic pipe piece grabbing method and shield tunneling machine

Abstract

The invention discloses an automatic duct piece grabbing method and a shield tunneling machine. The problems that in the prior art, duct piece grabbing efficiency is low, and safety is not high are solved. The duct piece automatic grabbing method comprises the steps that S1, hardware layout is conducted; s2, calibrating an image acquisition range of the safely grabbed segment; s3, converting the laser radar and the erector coordinate system; s4, identifying the position posture of the duct piece by the laser radar; s5, calculating to obtain the position posture of the duct piece relative to the erector through coordinate system conversion; s6, the erector is controlled to move to the position above the duct piece according to the position posture of the duct piece relative to the erector, the image collecting device collects the geometrical characteristics of the surface of the duct piece, and the posture of the duct piece relative to the erector is rechecked. The relative poses of the erector and the duct pieces are checked again through visual features, and grabbing safety is guaranteed; compared with an existing automatic assembly scheme in a single sensing mode, the method has the advantages of being rapid and safe.

Description

Technical Field

[0001] This invention relates to the field of tunnel boring machine technology, and in particular to a method for grabbing tunnel segments. Background Technology

[0002] As core equipment in underground engineering, the construction efficiency and quality of the tunnel directly depend on the precision and reliability of segment assembly. Segment assembly is a crucial process in shield tunneling, mainly including segment grabbing, spatial delivery, posture adjustment, and positioning installation. In traditional operation modes, these steps heavily rely on the operator's experience and manual control, resulting in high labor intensity, harsh working environments (high dust, high humidity, strong vibration), and the accumulation of human error.

[0003] In recent years, with the development of intelligent construction technology, domestic and foreign tunnel boring machine (TBM) manufacturers and research institutes have gradually carried out research and development of automatic segment assembly systems, attempting to replace manual operation with automation and intelligent means to improve construction efficiency, ensure operational safety, and improve construction quality.

[0004] To overcome the shortcomings of manual assembly, some research institutions have proposed automated assembly schemes based on a single sensing mode, such as the automatic grasping method and system for a shield tunnel segment assembly machine using a vision platform (publication number CN107152295B). This type of scheme uses industrial cameras to acquire images of the tunnel segments, identifies segment features through visual algorithms, and calculates relative pose. However, the single sensing mode lacks redundancy verification, making it prone to positioning errors under complex working conditions, leading to safety hazards such as segment collisions and detachment. Therefore, how to significantly improve assembly efficiency while ensuring operational safety is a key technical problem that urgently needs to be solved in the field of automated tunnel segment assembly. Summary of the Invention

[0005] To address the shortcomings in the aforementioned background technology, this invention proposes an automatic segment grabbing method and a tunnel boring machine, which solves the problems of low segment grabbing efficiency and low safety in the prior art.

[0006] The technical solution of the present invention is implemented as follows: an automatic segment grabbing method, the steps of which are: S1: hardware deployment: a laser radar is deployed under the traveling beam of the assembly machine so that the working range of the laser radar can cover the segments on the segment trolley located in the grabbing area; an image acquisition device is deployed on the grabbing head of the assembly machine so that the field of view of the image acquisition device can cover the geometric features of the segment surface.

[0007] S2: Image acquisition range calibration for safe segment grabbing: Control the grabbing head of the assembly machine to reach directly above the segment, record the initial position of the segment surface geometric features in the image of the image acquisition device, move the grabbing head back and forth to the maximum allowable grabbing deviation position between the segment and the grabbing head designed, and record the position of the segment surface geometric features in the image of the image acquisition device as the deviation boundary for safe segment grabbing, and determine the safe grabbing range.

[0008] S3: Coordinate system transformation between lidar and assembly machine: Attach reflective targets to the lifting heads of both the lidar and the assembly machine; measure the positions of the two reflective targets using a total station. , The position of the lidar relative to the coordinate system of the assembly machine .

[0009] S4: LiDAR identifies the position and attitude of the pipe segment. After the tunnel segment is moved into position by the tunnel segment trolley, the lidar scans the surface of the segment to obtain a point cloud of the target segment. Using a point cloud matching algorithm, the position and attitude of the segment relative to the lidar are calculated. .

[0010] S5: Calculate the position and orientation of the tunnel segment relative to the assembly machine through coordinate system transformation. .

[0011] S6: Based on the position and posture of the segment relative to the assembly machine, control the assembly machine to move above the segment, the image acquisition device collects the geometric features of the segment surface, and re-verifies the position and posture of the segment relative to the assembly machine; when the segment is within the safe range, the grab head drops to grab the segment.

[0012] This method utilizes the recognition function of LiDAR to achieve rapid positioning and reach above the tunnel segment in a single operation. Furthermore, it uses visual features to re-verify the relative pose between the assembly machine and the tunnel segment, ensuring safe handling. Compared to existing automated assembly schemes using a single sensor mode, this method offers the dual advantages of speed and safety.

[0013] In step S1, the lidar is kept relatively fixed to the traveling beam of the assembly machine; and the lidar corresponds to the assembly machine and the segment trolley located in the gripping area.

[0014] In step S2, the geometric features of the tunnel segment surface are pre-fabricated slots or corners on the segment surface. This ensures that the image acquisition device, such as a camera, accurately identifies the segment features.

[0015] The specific process of calculating the position and attitude of the pipe segment relative to the lidar using the point cloud matching algorithm in step S4 is as follows:

[0016] The target pipe segment point cloud acquired by the S4.1 lidar is When the segment is at the ideal grasping point, the original point cloud of the segment is: Based on the known quantity of the target pipe segment point cloud and the original point cloud of the pipe segment Solving for the rigid body transformation matrix yields the solution. ,in It is a rotation transformation matrix. It is a translation vector.

[0017] S4.2 Based on the rigid body transformation matrix This yields a rough position and orientation of the tube segment relative to the lidar. .

[0018] S4.3 Rotation Transformation Matrix Translation vector Perform iterative updates to output a high-precision rotation transformation matrix. Then the position and attitude of the tube segment relative to the lidar .

[0019] In step S4.3, the initial pose estimation is first set according to the position of the tunnel segment relative to the tunnel boring machine. , , Then, the iterative nearest-point method is used for local optimization to obtain a high-precision pose. and .

[0020] The coordinate system transformation in step S5 is as follows:

[0021] .

[0022] Initial pose estimation The specific process of using the iterative nearest point method for local optimization is as follows: For each point in the original point cloud of the pipe segment... , In the target pipe segment point cloud Find the nearest point To form a response Construct a least-squares objective function; minimize the sum of squared distances between the objective point and its corresponding points, then center and use SVD to solve for the optimal R,t; then iteratively update until the transformation increment is less than a threshold, the objective function decreases less than a threshold, and the maximum number of iterations is reached; finally, output a high-precision value. .

[0023] In step S6, the image acquisition device acquires the geometric features of the tube segment surface and verifies whether the position of the geometric features of the tube segment surface in the image is within the grabbing safety range defined in step S2. When the geometric features of the tube segment surface are within the defined safety range, the grabbing head falls to grab the tube segment; otherwise, the grabbing head is not allowed to fall.

[0024] The lidar uses a planar array lidar. The image acquisition device is an industrial planar array camera.

[0025] A tunnel boring machine includes an assembly machine and an assembly machine traveling beam mounted on the shield body. The assembly machine can move axially along the assembly machine traveling beam. During operation, the aforementioned automatic segment grabbing method is used to grab segments.

[0026] The beneficial effects of this invention are as follows: This invention utilizes the recognition function of lidar to achieve one-time positioning and quickly reach the top of the tube segment. In addition, the relative posture of the assembly machine and the tube segment can be re-verified through visual features to ensure the safety of grasping. Compared with the existing automatic assembly scheme with a single sensing mode, this method has the dual advantages of speed and safety.

[0027] This invention utilizes a lidar system to cover the segment-grabbing area, thereby identifying the segment's position and orientation relative to the lidar. Based on the relative orientation of the lidar and the segment-assembly machine, the relative orientation of the segment relative to the assembly machine is calculated. Through inverse kinematics calculations of the assembly machine, the target grabbing position of the machine's cylinders and motors is obtained, controlling the assembly machine's grabbing head to move above the segment. By installing an image acquisition device, such as a camera, on the assembly machine's grabbing head, the image position of the segment in the camera is checked again to ensure it matches the calibrated position, further improving safety and construction efficiency. This invention constructs a closed-loop control system for segment grabbing—a system of perception-decision-execution-confirmation—through a three-layer architecture of global lidar perception, precise kinematic solution, and local visual verification. This system offers significant technical advantages in accuracy, safety, and intelligence, further improving the construction efficiency of tunnel boring machines. Attached Figure Description

[0028] To more clearly illustrate the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0029] Figure 1 This is a schematic diagram of the control flow of the present invention;

[0030] Figure 2 This is a schematic diagram of the arrangement of the lidar and image acquisition device of the present invention. Detailed Implementation

[0031] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0032] Example 1, such as Figure 1 , 2 As shown, an automatic segment grasping method includes the following steps: S1: Hardware deployment: A LiDAR 3 is deployed under the traveling beam 1 of the assembly machine, ensuring that the LiDAR's operating range covers the segment 5 on the segment trolley 6 located in the grasping area; an image acquisition device 4 is deployed on the grasping head of the assembly machine, ensuring that the field of view of the image acquisition device 4 covers the geometric features of the segment surface. The geometric features of the segment surface include surface flatness, curvature accuracy, splicing seam shape, and bolt hole / grouting hole positions. The LiDAR is kept relatively fixed to the traveling beam of the assembly machine; and the LiDAR corresponds to the assembly machine and the segment trolley located in the grasping area; the LiDAR can identify the segment using an array LiDAR, a structured light-based camera, or a binocular vision-based 3D camera as an alternative.

[0033] S2: Image Acquisition Range Calibration for Safe Segment Grabbing: The assembly machine's gripping head is positioned directly above the segment. The initial position of the segment's surface geometric features in the image acquisition device 4 is recorded. The gripping head is moved back and forth to the maximum allowable gripping deviation position between the segment and the gripping head, and the position of the segment's surface geometric features in the image acquisition device 4 is recorded as the deviation boundary for safe segment grabbing, thus determining the safe grabbing range. In this embodiment, the segment's surface geometric features are pre-fabricated slots or corners on the segment's surface; that is, these features can be pre-fabricated on the segment's surface, such as circular, square, or triangular slots, or the corners of the segment can be identified as an alternative.

[0034] S3: Coordinate system transformation between lidar and assembly machine: Attach reflective targets to the lifting heads of both the lidar and the assembly machine; measure the positions of the two reflective targets using a total station. , The position of the lidar relative to the coordinate system of the assembly machine The assembly machine can move axially along the traveling beam; the total station uses phase-type or pulse-type laser ranging, combined with an absolute coding dial, to eliminate mechanical wear errors; the reflective characteristics of the reflective target ensure stable signal strength and reduce random measurement errors.

[0035] S4: LiDAR identifies the position and attitude of the pipe segment. After the tunnel segment is moved into position by the tunnel segment trolley, the lidar scans the surface of the segment to obtain a point cloud of the target segment. Using a point cloud matching algorithm, the position and attitude of the segment relative to the lidar are calculated. .

[0036] S5: Calculate the position and orientation of the tunnel segment relative to the assembly machine through coordinate system transformation. The coordinate system transformation is as follows:

[0037] .

[0038] S6: Based on the position and posture of the segment relative to the assembly machine, control the assembly machine to move above the segment, and the image acquisition device (4) collects the geometric features of the segment surface and rechecks the position and posture of the segment relative to the assembly machine; when the segment is within the safe range, the grab head falls to grab the segment.

[0039] As described above, the lidar can cover the segment-grabbing area, thereby identifying the segment's position and orientation relative to the lidar. Based on the relative orientation of the lidar and the segment assembly machine, the relative orientation of the segment relative to the assembly machine is calculated. Using the inverse kinematics calculation method of the assembly machine, the target grabbing position of the assembly machine's cylinders and motors is obtained, controlling the assembly machine's grabbing head to move above the segment. However, since lidar detection may have deviations, and accidental human contact could alter the lidar's coordinates, affecting the relative orientation calculation between the segment and the assembly machine, an image acquisition device, such as a camera, is installed on the assembly machine's grabbing head to ensure safety. This camera checks whether the segment's position in the camera's image matches the calibrated position, or whether the deviation is within acceptable limits. If the deviation exceeds the acceptable limit, the grabbing operation is not allowed.

[0040] Example 2: An automatic segment acquisition method, further optimized based on Example 1. In this example, the lidar 3 is a planar lidar; the image acquisition device 4 is an industrial planar camera. The specific process of calculating the segment's position and attitude relative to the lidar using a point cloud matching algorithm in step S4 of this example is as follows: S4.1 The target segment point cloud acquired by the lidar is... When the segment is at the ideal grasping point, the original point cloud of the segment is: The target pipe segment point cloud is The original point cloud of the tunnel segment is All data are known, based on the target segment point cloud. and the original point cloud of the pipe segment Solving for the rigid body transformation matrix yields the solution. ,in It is a rotation transformation matrix. It is a translation vector; making the transformed point cloud , and the target point cloud To overlap as much as possible in space, among which These are the x, y, z coordinates of each point in the original point cloud of the pipe segment, which can be represented as: .

[0041] S4.2 Based on the rigid body transformation matrix This yields a rough position and orientation of the tube segment relative to the lidar. .

[0042] S4.3 Rotation Transformation Matrix Translation vector Perform iterative updates to output a high-precision rotation transformation matrix. Then the position and attitude of the tube segment relative to the lidar .

[0043] In step S4.3, the initial pose estimation is first set according to the position of the tunnel segment relative to the tunnel boring machine. The position of the tunnel segment relative to the tunnel boring machine is based on a center value that fluctuates within a range. Coarse registration is set using this center value as a reference. This estimation does not require precision and can be set to... , Then, the iterative nearest-point method is used for local optimization to obtain a high-precision pose. and Based on the initial pose given by coarse registration, the Iterative Closest Point (ICP) method is used for local optimization to obtain a high-precision pose. Specifically, for each point in the original point cloud of the pipe segment... , In the target pipe segment point cloud Find the nearest Q in the middle. To form a response Construct a least-squares objective function; minimize the sum of squared distances between the target point and its corresponding points, using the formula... ; yes The number of points in a point cloud Indicates the first step in the calculation process There are several points. Then, centering and SVD are used to solve for the optimal R,t. The process involves calculating the centroids of the source point cloud and the target point cloud. , , and They represent and The coordinate center of all points in the point cloud; centered calculation, the result is obtained. , Calculate the covariance matrix For the matrix SVD decomposition ,in Describes a left singular matrix. Describes a right singular matrix. Represent the singular value matrix; calculate the rotation matrix. ,like Then flip a column, in which Represents the computation matrix Determinant: ,in This represents a diagonal matrix with diagonal elements of 1, 1, -1; the translation vector is calculated to obtain... Then iterates and updates until the transformation increment is less than a threshold, the objective function decreases less than a threshold, and the maximum number of iterations is reached; finally, a high-precision output is obtained. That is, the precise position and pose of the point cloud. .

[0044] In step S6, the image acquisition device 4 acquires the geometric features of the pipe segment surface and verifies whether the position of these features in the image is within the safe grasping range defined in step S2. If the features are within the defined safe range, the grasping head drops to grasp the pipe segment; otherwise, the grasping head is not allowed to drop. The lidar covers the pipe segment; the camera's field of view covers the features of the pipe segment; by calculating the position of the features in the image, the deviation between the current feature position and the calibrated position can be compared to detect whether the position and orientation of the pipe segment and the grasping head meet the grasping requirements. If they do, the segment is grasped; otherwise, it is not. This method, compared to existing methods that either use lidar or rely solely on a camera, has the dual advantages of speed and safety.

[0045] Example 3: A tunnel boring machine (TBM) includes an assembly machine and an assembly machine travel beam mounted on the shield body. The assembly machine can move axially along the assembly machine travel beam. The axial movement of the travel beam allows the assembly machine to grab segments at any position without waiting for the shield body to adjust its attitude, forming a parallel workflow with the TBM propulsion system and eliminating waiting time between processes. During operation, the aforementioned automatic segment grabbing method is used for segment grabbing. The operation process is as follows: a segment transport trolley delivers the segment to a designated position at the bottom of the shield tail; the assembly machine moves axially along the travel beam to the grabbing position; the camera range for safe segment grabbing is calibrated; a lidar identifies the segment's posture; and then the segment's posture is converted to the assembly machine coordinate system; the segment grabbing motion target is calculated; based on the segment's position and posture relative to the assembly machine, the assembly machine moves above the segment; the camera collects segment features; the deviation between the features and the calibrated features is calculated; then the deviation is checked; if the conditions are met, the assembly machine's grabbing head drops to grab the segment; thus, automatic segment grabbing is achieved.

[0046] This invention effectively and quickly identifies the relative position of the tunnel segment and the assembly machine, guiding the assembly machine's gripping head to reach above the segment in a single movement. Upon reaching the segment, a visual camera performs a safety check, ensuring the safety of the gripping process. This solution effectively meets the speed and safety requirements of automated tunnel segment gripping in practical applications, significantly improving the efficiency and quality of tunnel segment assembly.

[0047] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An automatic segment picking method, characterized in that: Step S1: Hardware deployment: A lidar is deployed under the traveling beam of the assembly machine so that the working range of the lidar can cover the segments on the segment trolley in the grabbing area; an image acquisition device is deployed on the grabbing head of the assembly machine so that the field of view of the image acquisition device can cover the geometric features of the segment surface. S2: Image acquisition range calibration for safe segment grabbing: Control the grabbing head of the assembly machine to reach directly above the segment, record the initial position of the segment surface geometric features in the image of the image acquisition device, move the grabbing head back and forth to the maximum allowable grabbing deviation position between the segment and the grabbing head, and record the position of the segment surface geometric features in the image of the image acquisition device as the deviation boundary for safe segment grabbing, and determine the safe grabbing range. S3: Coordinate system transformation between lidar and assembly machine: Attach reflective targets to the lifting heads of both the lidar and the assembly machine; measure the positions of the two reflective targets using a total station. , The position of the lidar relative to the coordinate system of the assembly machine ; S4: LiDAR identifies the position and attitude of the pipe segment. After the tunnel segment is moved into position by the tunnel segment trolley, the lidar scans the surface of the segment to obtain a point cloud of the target segment. Using a point cloud matching algorithm, the position and attitude of the segment relative to the lidar are calculated. ; S5: Calculate the position and orientation of the tunnel segment relative to the assembly machine through coordinate system transformation. ; S6: Based on the position and posture of the segment relative to the assembly machine, control the assembly machine to move above the segment, the image acquisition device collects the geometric features of the segment surface, and re-verifies the position and posture of the segment relative to the assembly machine; when the segment is within the safe range, the grab head drops to grab the segment.

2. The automatic segment picking method according to claim 1, characterized in that: In step S1, the lidar is kept relatively fixed to the traveling beam of the assembly machine; and the lidar corresponds to the assembly machine and the segment trolley located in the gripping area.

3. The automatic segment picking method according to claim 1, characterized in that: In step S2, the geometric features of the segment surface are pre-made slots or corners on the segment surface.

4. The automatic segment picking method according to any one of claims 1 to 3, characterized in that: The specific process of calculating the position and attitude of the pipe segment relative to the lidar using the point cloud matching algorithm in step S4 is as follows: The target pipe segment point cloud acquired by the S4.1 lidar is When the segment is at the ideal grasping point, the original point cloud of the segment is: Based on the target segment point cloud and the original point cloud of the pipe segment Solving for the rigid body transformation matrix yields the solution. ,in It is a rotation transformation matrix. It is a translation vector; S4.2 Based on the rigid body transformation matrix This yields a rough position and orientation of the tube segment relative to the lidar. ; S4.3 Rotation Transformation Matrix Translation vector Perform iterative updates to output a high-precision rotation transformation matrix. Then the position and attitude of the tube segment relative to the lidar .

5. The automatic segment picking method according to claim 4, characterized in that: In step S4.3, the initial pose estimation is first set according to the position of the tunnel segment relative to the tunnel boring machine. , , Then, the iterative nearest-point method is used for local optimization to obtain a high-precision pose. and .

6. The automatic segment picking method according to claim 5, characterized in that: Initial pose estimation The specific process of using the iterative nearest point method for local optimization is as follows: For each point in the original point cloud of the pipe segment... , Point cloud of the target pipe segment Find the nearest point To form a response Construct a least-squares objective function; minimize the sum of squared distances between the objective point and its corresponding points, then center and use SVD to solve for the optimal R,t; then iteratively update until the transformation increment is less than a threshold, the objective function decreases less than a threshold, and the maximum number of iterations is reached; finally, output a high-precision value. .

7. The automatic segment grasping method according to claim 1 or 5, characterized in that: The coordinate system transformation in step S5 is as follows: .

8. The automatic segment picking method according to claim 7, characterized in that: In step S6, the image acquisition device acquires the geometric features of the tube segment surface and verifies whether the position of the geometric features of the tube segment surface in the image is within the grabbing safety range defined in step S2. When the geometric features of the tube segment surface are within the defined safety range, the grabbing head falls to grab the tube segment; otherwise, the grabbing head is not allowed to fall.

9. The automatic segment picking method according to claim 1, characterized in that: The lidar uses a surface array lidar; the image acquisition device is an industrial surface array camera.

10. A tunnel boring machine, characterized in that: It includes an assembly machine and an assembly machine traveling beam mounted on the shield body. The assembly machine can move axially along the assembly machine traveling beam. During operation, the automatic segment grabbing method described in any one of claims 1 to 7 is used to grab the segments.

Images