System and method for intelligently acquiring digital information of factory-end prefabricated duct piece based on structured light

Abstract

The invention provides a factory-end prefabricated segment digital information intelligent acquisition system and method based on structured light. The detection device comprises a supporting frame, an XYZ three-axis sliding module moving system, a two-dimensional rotating holder and a structured light camera. According to the supporting frame, the prefabricated duct pieces are arranged in the inner space through surrounding type design, and a stable non-contact scanning environment is formed. The top face of the supporting frame serves as an installation reference face of the XYZ three-axis sliding module moving system. The XYZ three-axis sliding module moving system comprises an X-axis sliding module I, a Y-axis sliding module, a Z-axis sliding module I, a Z-axis sliding module II and an X-axis sliding module II. The X-axis sliding module I and the X-axis sliding module II are arranged on the two sides of the prefabricated pipe piece in parallel and cover the length direction of the prefabricated pipe piece. The Y-axis sliding module stretches across the position between the X-axis sliding module I and the X-axis sliding module II and covers the width direction of the prefabricated pipe piece. The two ends of the Y-axis sliding module are slidably connected to the upper surface of the X-axis sliding module I and the upper surface of the X-axis sliding module II correspondingly. And the Z-axis sliding module I and the Z-axis sliding module II are vertically mounted on the side wall of the Y-axis sliding module. And the tail end of each group of Z-axis is provided with a two-dimensional rotating holder. And a structured light camera is mounted below the two-dimensional rotating holder. The double scanners can synchronously execute different-track scanning through three-axis cooperation in the frame, and are suitable for full-surface acquisition of complex curved surfaces.

Description

Technical Field

[0001] This invention relates to the field of industrial inspection technology, and in particular to a digital information intelligent acquisition system and method for precast tube segments at the factory end based on structured light. Background Technology

[0002] Precast tunnel segments serve as supporting and lining units for tunnels, underground engineering projects, or large structures, forming a stable overall structure through assembly. The geometric dimensions and surface quality inspection of precast tunnel segments are crucial for ensuring project safety and construction efficiency. Traditional inspection methods rely on manual measurement using tools such as measuring tapes and straightedges, which suffers from low efficiency and significant subjective errors. Due to their large size and curved surfaces, the geometric characteristics (such as curvature and flatness) of precast tunnel segments are difficult to measure accurately using conventional tools. Furthermore, the identification of surface quality defects (such as cracks, honeycombing, and pores) is highly dependent on worker experience, resulting in a high risk of missed defects.

[0003] In recent years, 3D laser scanning technology has been introduced into the field of industrial inspection due to its non-contact and high-precision characteristics. This technology can reconstruct 3D models of components from point cloud data, theoretically enabling full-surface coverage inspection. However, in practical applications, the complex curved surfaces of large precast tunnel segments mean that a single scan cannot acquire complete data, requiring multiple adjustments to the scanning stations and multi-angle scans, which is time-consuming and complex. The maximum stitching error of the point cloud from each station reaches 5mm. Meanwhile, while the introduction of high-precision equipment has improved inspection efficiency, the high procurement and maintenance costs limit its widespread adoption in small and medium-sized precast component factories.

[0004] In summary, developing a digital information intelligent acquisition system for precast tunnel segments at the factory level based on structured light is of great significance. Summary of the Invention

[0005] The purpose of this invention is to provide a digital information intelligent acquisition system and method for precast tunnel segments at the factory end based on structured light, so as to solve the problems existing in the prior art.

[0006] The technical solution adopted to achieve the purpose of this invention is as follows: a factory-end precast segment digital information intelligent acquisition system based on structured light, including a support frame, an XYZ three-axis sliding module moving system, a two-dimensional rotating gimbal and a structured light camera.

[0007] The support frame, through its enclosed design, places the prefabricated tube segments within its internal space, creating a stable, non-contact scanning environment. The top surface of the support frame serves as the mounting reference surface for the XYZ three-axis sliding module movement system.

[0008] The XYZ three-axis sliding module moving system includes an X-axis sliding module I, a Y-axis sliding module, a Z-axis sliding module I, a Z-axis sliding module II, and an X-axis sliding module II.

[0009] The X-axis sliding module I, Y-axis sliding module, Z-axis sliding module I, Z-axis sliding module II, and X-axis sliding module II all adopt a reciprocating rack and pinion slide structure. The reciprocating rack and pinion slide structure includes a guide rail and a slide that is slidably connected to the guide rail and can slide along the length of the guide rail. X-axis sliding modules I and II are arranged parallel to each other on both sides of the precast segment, covering the length direction of the precast segment. The Y-axis sliding module spans between X-axis sliding modules I and II, covering the width direction of the precast segment. The two ends of the guide rail of the Y-axis sliding module are respectively connected to the upper surface of the slide of X-axis sliding module I and X-axis sliding module II. The slides of Z-axis sliding modules I and Z-axis sliding module II are mounted on the side wall of the slide of the Y-axis sliding module. The guide rails of Z-axis sliding modules I and Z-axis sliding module II are arranged vertically. Each Z-axis sliding module is equipped with a two-dimensional rotating gimbal at its end, which can change the horizontal and vertical angles. A structured light camera is mounted under the two-dimensional rotating gimbal. The structured light camera can translate along the Y-axis and move vertically along the Z-axis.

[0010] Furthermore, the precast segments are placed on a placement frame. The placement frame includes a plurality of H-beams spaced apart along the X-axis. The lower flange of the H-beams is equipped with directional wheels, and the upper flange supports the precast segments.

[0011] Furthermore, it also includes a platform. The upper surface of the platform is provided with a slide rail for the insertion of directional wheels.

[0012] Furthermore, the X-axis sliding module I, Y-axis sliding module I, Z-axis sliding module II, and X-axis sliding module II are all controlled by a closed-loop stepper motor for linear motion.

[0013] The X-axis sliding module I and X-axis sliding module II include a base I and a slide I. A pair of guide rails are provided on the upper surface of the base I. The slide I is slidably mounted on the upper surface of the base I via the guide rails. A rack located between the pair of guide rails is also provided on the upper surface of the base I. The length direction of the rack is parallel to the length direction of the guide rails. A mounting through hole is provided on the upper surface of the slide I. A motor mounting flange plate is fixedly installed in the mounting through hole. A drive motor is provided on the motor mounting flange plate. A gear is provided below the motor mounting flange plate, and the drive motor is connected to the gear. The gear meshes with the rack.

[0014] The Y-axis sliding module includes a base II and a slide II. A pair of guide rails are provided on the side wall of the base II. The slide II is slidably mounted on the side wall of the base II via the guide rails. A rack is also provided on the side wall of the base II between the pair of guide rails. The length direction of the rack is parallel to the length direction of the guide rails. A mounting through hole is provided on the side wall of the slide II. A motor mounting flange plate is fixedly installed in the mounting through hole. A drive motor is mounted on the motor mounting flange plate. A gear is provided beside the motor mounting flange plate, and the drive motor is connected to the gear. The gear meshes with the rack.

[0015] The Z-axis sliding module I and Z-axis sliding module II include a base III and a slide III. The base III is vertically arranged. The slide III is fixed on the slide II. A pair of guide rails are provided on the side wall of the base III near the slide II. The slide III is slidably mounted on the side wall of the base III via the guide rails. The side wall of the base III also has a rack located between the pair of guide rails. The length direction of the rack is parallel to the length direction of the guide rails. A mounting through hole is opened in the side wall of the slide III. A motor mounting flange plate is fixedly installed in the mounting through hole. A drive motor is provided on the motor mounting flange plate. A gear is provided on the side of the motor mounting flange plate, and the drive motor is connected to the gear. The gear meshes with the rack.

[0016] Furthermore, the support frame is constructed from horizontal and vertical aluminum alloy profiles joined together to form an open hexahedral structure. An infrared sensor array is integrated into the inner wall of the support frame.

[0017] The technical effects of this invention are beyond doubt:

[0018] A. The dual scanners can perform cross-track scanning simultaneously through three-axis collaboration within the frame, making them suitable for full-surface acquisition of complex curved surfaces;

[0019] B. The support frame achieves high strength support through lightweight aluminum alloy material, while avoiding electromagnetic interference issues; the width and height of the frame are expandable, and the scanning coverage of precast segments is achieved through spliced ​​aluminum alloy profiles. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the device structure;

[0021] Figure 2 This is a schematic diagram illustrating the principle of structured light scanning measurement.

[0022] Figure 3 This is a schematic diagram of the device's operation.

[0023] Figure 4 Flowchart of a factory-end precast segment digital information intelligent acquisition system based on structured light;

[0024] Figure 5The point cloud image of a precast segment was not registered after multiple scans;

[0025] Figure 6 This is a coarsely stitched cloud map of the point clouds of multiple precast tunnel segments after homogeneous coordinate transformation.

[0026] Figure 7 The overall point cloud image of the precast segment after accurate registration using the Generalized-ICP algorithm.

[0027] In the diagram: 1-Aluminum alloy support frame, 2-X-axis sliding module I, 3-Y-axis sliding module, 4-Z-axis sliding module I, 5-Z-axis sliding module II, 6-Two-dimensional rotating gimbal, 7-Structured light camera, 8-X-axis sliding module II, 9-86 stepper motor I driving the Y-axis sliding module along the X-axis, 10-86 stepper motor driving the Z-axis sliding module I along the Y-axis, 11-86 stepper motor driving the Z-axis sliding module II along the Y-axis, 12-86 stepper motor II driving the Y-axis sliding module along the X-axis, 13-86 stepper motor driving the Z-axis sliding module I to move up and down, 14-86 stepper motor driving the Z-axis sliding module II to move up and down, 15-Precast tube segment, 16-H-beam, 17-Directional wheel. Detailed Implementation

[0028] The present invention will be further described below with reference to embodiments, but it should not be construed that the scope of the present invention is limited to the following embodiments. Various substitutions and modifications made based on ordinary technical knowledge and common practices in the art without departing from the above-described technical concept of the present invention should be included within the scope of protection of the present invention.

[0029] Example 1:

[0030] See Figure 2 The scanning and measurement principle of a surface structured light camera is based on structured light technology. By projecting light of a known pattern (such as stripes or grids) onto the surface of an object, the light pattern deforms according to the shape of the object's surface. After the camera captures these deformed light patterns, it uses image processing and analysis, and calculates the three-dimensional coordinates of each point using the principle of triangulation. Specifically, the geometry of the object's surface determines the offset and deformation of the light pattern. By combining the positional relationship between the light source and the camera, and the degree of pattern deformation, the device can accurately calculate the position of each point on the surface.

[0031] See Figure 1 To address the issues of data redundancy, incomplete coverage of complex curved surfaces, and excessive manual intervention in existing technologies, this embodiment provides a factory-end precast segment digital information intelligent acquisition system based on structured light, including a support frame 1, an XYZ three-axis sliding module moving system, a two-dimensional rotating gimbal 6, and a structured light camera 7.

[0032] The support frame 1, through its enclosed design, places the prefabricated tube segment 15 within its internal space, creating a stable non-contact scanning environment. The top surface of the support frame 1 serves as the mounting reference surface for the XYZ three-axis sliding module moving system.

[0033] The XYZ three-axis sliding module moving system includes an X-axis sliding module I2, a Y-axis sliding module 3, a Z-axis sliding module I4, a Z-axis sliding module II5, and an X-axis sliding module II8. The X-axis sliding modules I2 and II8 are arranged parallel to each other on both sides of the precast segment 15, covering the length direction of the precast segment 15. The Y-axis sliding module 3 spans between the X-axis sliding modules I2 and II8, covering the width direction of the precast segment 15. The two ends of the Y-axis sliding module 3 are slidably connected to the upper surfaces of the X-axis sliding modules I2 and II8, respectively. The Z-axis sliding modules I4 and II5 are vertically mounted on the sidewalls of the Y-axis sliding module 3. Each Z-axis is equipped with a two-dimensional rotating gimbal 6. A structured light camera 7 is mounted under the two-dimensional rotating gimbal 6.

[0034] Example 2:

[0035] This embodiment is similar in main content to Embodiment 1, but also includes a platform. The precast tunnel segment 15 is placed on a placement frame. The placement frame includes a plurality of H-beams 16 spaced apart along the X-axis. The lower flange of each H-beam 16 is equipped with a directional wheel 17, and the upper flange supports the precast tunnel segment 15. A slide rail for the directional wheel 17 to be inserted is provided on the upper surface of the platform. The scaled-down model 15 of the tower tunnel segment is placed on the tunnel segment placement frame composed of 500×300 HN-shaped steel and the directional wheel 17. The tunnel segment placement frame can move along its major axis (X-axis). Figure 2 As shown, the precast segments are placed on the segment placement rack, ensuring that the rack does not obstruct the holes on the bottom of the segments.

[0036] Example 3:

[0037] The main content of this embodiment is the same as that of embodiment 1 or 2, wherein, see [link / reference]. Figure 3 The X-axis sliding module I2, Y-axis sliding module 3, Z-axis sliding module I4, Z-axis sliding module II5 and X-axis sliding module II8 are all driven by closed-loop stepper motors.

[0038] The X-axis sliding module I2, Y-axis sliding module 3, Z-axis sliding module I4, Z-axis sliding module II5 and X-axis sliding module II8 are all controlled by a closed-loop stepper motor for linear motion.

[0039] The X-axis sliding module I2 and X-axis sliding module II8 include a base I and a slide I. A pair of guide rails are provided on the upper surface of the base I. The slide I is slidably mounted on the upper surface of the base I via the guide rails. A rack located between the pair of guide rails is also provided on the upper surface of the base I. The length direction of the rack is parallel to the length direction of the guide rails. A mounting through hole is provided on the upper surface of the slide I. A motor mounting flange plate is fixedly installed in the mounting through hole. A drive motor is provided on the motor mounting flange plate. A gear is provided below the motor mounting flange plate, and the drive motor is connected to the gear. The gear meshes with the rack.

[0040] The Y-axis sliding module 3 includes a base II and a slide II. A pair of guide rails are provided on the side wall of the base II. The slide II is slidably mounted on the side wall of the base II via the guide rails. A rack is also provided on the side wall of the base II between the pair of guide rails. The length direction of the rack is parallel to the length direction of the guide rails. A mounting through hole is provided on the side wall of the slide II. A motor mounting flange plate is fixedly installed in the mounting through hole. A drive motor is mounted on the motor mounting flange plate. A gear is provided beside the motor mounting flange plate, and the drive motor is connected to the gear. The gear meshes with the rack.

[0041] The Z-axis sliding module I4 and Z-axis sliding module II5 include a base III and a slide III. The base III is vertically arranged. The slide III is fixed on the slide II. A pair of guide rails are provided on the side wall of the base III near the slide II. The slide III is slidably mounted on the side wall of the base III via the guide rails. The side wall of the base III also has a rack located between the pair of guide rails. The length direction of the rack is parallel to the length direction of the guide rails. A mounting through hole is opened on the side wall of the slide III. A motor mounting flange plate is fixedly installed in the mounting through hole. A drive motor is provided on the motor mounting flange plate. A gear is provided next to the motor mounting flange plate, and the drive motor is connected to the gear. The gear meshes with the rack.

[0042] It is worth noting that the 86-stepper motor driving the Y-axis sliding module along the X-axis is marked as 9. The 86-stepper motor driving Z-axis sliding module I along the Y-axis is marked as 10. The 86-stepper motor driving Z-axis sliding module II along the Y-axis is marked as 11. The 86-stepper motor driving the Y-axis sliding module along the X-axis is marked as 12. The 86-stepper motor driving Z-axis sliding module I to move up and down is marked as 13. The 86-stepper motor driving Z-axis sliding module II to move up and down is marked as 14.

[0043] Example 4:

[0044] The main content of this embodiment is the same as any one of embodiments 1 to 3, wherein the support frame 1 adopts an open hexahedral structure spliced ​​from horizontal and vertical aluminum alloy profiles. The inner wall of the support frame 1 integrates an infrared sensor array, which triggers emergency braking when the moving trajectory of the structured light camera 7 approaches the surface of the precast tube segment 15 to avoid mechanical collision.

[0045] Example 5:

[0046] See Figure 4 This embodiment provides a data acquisition method for a factory-end precast segment digital information intelligent acquisition system based on structured light, as described in any one of Embodiments 1 to 4, comprising the following steps:

[0047] 1) Place the precast segment 15 at the acquisition position. Control the closed-loop stepper motor movement through PLC to move the sliding module XYZ three axes to the point cloud data acquisition points on both sides of the precast segment.

[0048] 2) Adjust the horizontal and vertical angles of the 2D rotating gimbal 6 via the host computer to ensure that the structured light cameras 7 on both sides of the precast tube segment 15 are accurately positioned to be scanned. The horizontal rotation range of the 2D gimbal is -180° to 180°, and its vertical rotation range is -90° to 90°. Use software to control the host computer to send control signals to control the accurate horizontal and vertical rotation of the 2D gimbal, so that the 3D structured light scanner mounted on the 2D gimbal can accurately position the precast tube segment.

[0049] 3) Use the structured light camera 7 to scan the precast tube segment 15 and acquire single-frame point cloud data. The motion control module receives the input three-dimensional spatial pose information, enabling the two 3D structured light scanners to accurately reach the target position and precisely locate the precast tube segment to be scanned. After adjusting the spatial pose of the 3D structured light scanners, the computer controls the scanners to acquire single-frame point cloud data of the precast tube segment.

[0050] 4) Statistical filtering algorithms are used to denoise the point cloud data, removing sparsely distributed outliers in space. Point cloud data is acquired using a 3D structured light scanner. The point cloud data may contain noise, and may also be distorted or contain artifacts due to motion instability. Statistical filtering algorithms are used to statistically analyze the neighborhood of each point in the captured single-frame point cloud. Based on the distance distribution characteristics from the point to all neighboring points, outliers that are sparsely distributed and far from dense areas are removed. In this embodiment, statistical filtering algorithms are used to remove points whose average distance from their neighborhood exceeds a distance threshold. A parameter for the number of points within the neighborhood is set, and the distance threshold thresh_d = μ + 3σ is set to complete the outlier removal. Step 4) specifically includes the following sub-steps:

[0051] 4.1) Calculate the average distance from each point in the input point cloud to its k nearest neighbors. Obtain a Gaussian distribution of the average distance of the k nearest neighbors by statistics, and calculate the mean μ and standard deviation σ of the Gaussian distribution to determine the distance threshold thresh_d, thresh_d = μ + k·σ (k = 1, 2, 3).

[0052] 4.2) Based on the distance threshold determined in the first step, traverse each point in the input point cloud. If the average distance of the k neighboring points of a point is greater than the distance threshold thresh_d, then the point is identified as an outlier and deleted, thus achieving the noise reduction processing of the input point cloud.

[0053] 5) Use a voxel downsampling algorithm to reduce the point cloud density and obtain downsampled single-frame point cloud data. The point cloud data acquired by the 3D structured light scanner is large, so a voxel downsampling algorithm is used to lightweight the data. The voxel downsampling algorithm divides the 3D space into a uniformly sized voxel grid. For each point within a voxel, the algorithm generates a "representative point" by taking the centroid, mean, or median of all points within the voxel, and then removes all other points within the voxel. In step 5), based on the 3D coordinate set of the point cloud data, the maximum value x on the x, y, and z coordinate axes is calculated. max y max z max and minimum value x min y min z min Design the side length r of the voxel mesh, based on x. max y max z max x min y min z min Calculate the side length l of the minimum bounding box of the point cloud. x l y l z :

[0054]

[0055] Calculate the size of the voxel mesh.

[0056]

[0057] In the formula, This indicates rounding down to the nearest integer.

[0058] Calculate the index h of each point in the point cloud within the voxel grid.

[0059]

[0060] In the formula, h x hy h z represents the 3D index of the voxel mesh, and h represents the linear index of the voxel mesh.

[0061] Store the linear index h of each point in an array in ascending order for later access. Then, based on the 3D index h of each point... x h y h z Each point is assigned to a corresponding voxel grid, and the centroid of each voxel grid is calculated. The centroid is then used to replace all points within the grid. This results in the voxel-downsampled point cloud data, achieving lightweight point cloud data. The single-frame scan point cloud image of the precast tunnel segment in the implementation example is shown below. Figure 5 As shown.

[0062] 6) The translation data of each sliding module and the rotation angle of the 2D gimbal are acquired through a three-axis movable rotatable motion control module. Then, by calculating the homogeneous coordinate transformation matrix between two adjacent frames of point clouds from the two structured light cameras, coarse stitching between adjacent point clouds is achieved. In step 6), the rotation matrix around the x-axis of the 3D structured light scanner coordinate system is calculated based on the pitch rotation angle α of the 2D gimbal. Calculate the rotation matrix around the y-axis of the 3D structured light scanner based on the horizontal rotation angle β of the 2D gimbal.

[0063]

[0064] Using the above rotation matrix and The coordinates P of the second point cloud in the 3D structured light scanner coordinate system are... A Transform to the coordinates P of the first point cloud in the 3D structured light scanner coordinate system B .

[0065]

[0066] By calculating the displacement changes t = (Δx, Δy, Δz) of the XYZ three-axis sliding module in the x, y, and z axes of the 3D structured light scanner coordinate system, and combining this with the aforementioned rotation matrix, an augmented matrix is ​​formed. Adding a row vector (0,0,0,1) to the augmented matrix yields the homogeneous coordinate transformation matrix between two adjacent point clouds, thus realizing the coordinate transformation of the two point clouds in the overall device coordinate system:

[0067]

[0068] The coordinates of two adjacent point clouds are transformed to the overall coordinate system specified by the scanning system using a homogeneous coordinate transformation matrix, achieving coarse stitching between adjacent point clouds. The stitching quality is affected by the accuracy of the device. In the implementation example, the coarsely stitched point cloud of a single frame of precast tunnel segment after homogeneous coordinate transformation is shown below. Figure 6 As shown.

[0069] 7) The GICP (Generalized Iterative Closest Point) algorithm is used to accurately register two adjacent point clouds. The above steps are repeated to obtain the overall point cloud data for both sides of the precast segment. Finally, the GICP algorithm is used again to accurately register the two point clouds, obtaining the overall point cloud data for the surface of the precast segment. After coarse registration, the Generalized ICP algorithm is used to accurately register two adjacent point clouds, and the above steps are repeated to finally obtain the overall point cloud of the precast segment. The core of the GICP algorithm is to incorporate the local geometric characteristics of the point cloud (e.g., the local plane, surface, or spatial distribution of the points) into the registration process. Specifically, the basic principle of the GICP registration algorithm is to combine the "point-to-point" ICP algorithm and the "point-to-plane" ICP algorithm. Its error function is based on two Gaussian models of their respective local distributions, using their covariance matrices to more accurately describe the matching error between the geometric structures. Therefore, the GICP algorithm has a wider range of applications than the standard ICP algorithm and can improve the accuracy and robustness of registration. Step 7) specifically includes the following sub-steps:

[0070] 7.1) Given two sets of point cloud data, including source point cloud P and target point cloud Q, initialize the rigid transformation matrix T.

[0071] 7.2) Calculate the covariance matrix. For each point p∈P and q∈Q, calculate the covariance matrices ΣP and ΣQ based on its neighborhood points:

[0072]

[0073] In the formula, μ P It is the centroid of a point within the domain of p.

[0074] 7.3) Use the nearest neighbor algorithm to find the nearest neighbor pair (p,q) of each point in the source point cloud P in the target point cloud Q.

[0075] 7.4) Calculate the error function based on the Gaussian distribution form of the two points:

[0076]

[0077] In the formula, (p,q) are point pairs. ΣP and ΣQ are the covariance matrices of each point pair, describing their local geometric distribution. (ΣP+ΣQ) -1 It is the inverse of the total covariance matrix, used to measure the error between point pairs.

[0078] 7.5) Iteratively optimize the rigid transformation matrix T by minimizing the error function E:

[0079] T = argmin T E

[0080] 7.6) Apply the currently calculated rigid transformation matrix T to update the position of the source point cloud P, check the convergence condition of the error function, and if the error drops to a certain threshold or the number of iterations reaches a preset value, the accurate registration of the two frame point clouds is finally achieved.

[0081] In this embodiment, the overall point cloud after the coarse splicing point cloud of the precast tunnel segments is precisely registered using the Generalized-ICP algorithm is as follows: Figure 7 As shown in the figure. Experimental results demonstrate that this embodiment can complete intelligent scanning of precast segments, can use algorithms to achieve coarse stitching and registration of point clouds between adjacent frames, can use algorithms to achieve precise registration of coarse stitched point clouds, and can complete intelligent scanning of the overall point cloud of precast segments. The method described in this invention is truly effective.

Claims

1. A factory-end precast tunnel segment digital information intelligent acquisition system based on structured light, characterized in that: It includes a support frame (1), an XYZ three-axis sliding module moving system, a two-dimensional rotating gimbal (6), and a structured light camera (7); The support frame (1) uses an enclosed design to place the prefabricated tube segment (15) in the internal space to form a stable non-contact scanning environment; the top surface of the support frame (1) serves as the installation reference surface for the XYZ three-axis sliding module moving system. The XYZ three-axis sliding module moving system includes an X-axis sliding module I (2), a Y-axis sliding module (3), a Z-axis sliding module I (4), a Z-axis sliding module II (5), and an X-axis sliding module II (8); The X-axis sliding module I (2), Y-axis sliding module (3), Z-axis sliding module I (4), Z-axis sliding module II (5) and X-axis sliding module II (8) all adopt a reciprocating rack and pinion slide structure; the reciprocating rack and pinion slide structure includes a guide rail and a slide that is slidably connected to the guide rail and can slide along the length direction of the guide rail; the X-axis sliding module I (2) and X-axis sliding module II (8) are arranged in parallel on both sides of the precast tube segment (15) and cover the length direction of the precast tube segment (15); the Y-axis sliding module (3) spans between the X-axis sliding module I (2) and X-axis sliding module II (8) and covers the width direction of the precast tube segment (15); the two ends of the guide rail of the Y-axis sliding module (3) are respectively connected to the upper surface of the slide of the X-axis sliding module I (2) and X-axis sliding module II (8); The slides of the Z-axis sliding module I (4) and Z-axis sliding module II (5) are mounted on the side wall of the slide of the Y-axis sliding module (3); the guide rails of the Z-axis sliding module I (4) and Z-axis sliding module II (5) are arranged vertically; a two-dimensional rotating gimbal (6) is provided at the end of the guide rail, which can change the horizontal angle and the pitch angle; a structured light camera (7) is installed under the two-dimensional rotating gimbal (6); the structured light camera (7) can translate along the Y-axis and move vertically along the Z-axis.

2. The intelligent digital information acquisition system for precast tunnel segments at the factory end based on structured light according to claim 1, characterized in that: The precast tube segment (15) is placed on a placement frame; the placement frame includes a number of H-beams (16) spaced apart along the X-axis; the lower flange of the H-beams (16) is provided with a directional wheel (17), and the upper flange supports the precast tube segment (15).

3. The intelligent digital information acquisition system for precast tunnel segments at the factory end based on structured light according to claim 2, characterized in that: It also includes a platform; the upper surface of the platform is provided with a slide rail for the directional wheel (17) to be embedded.

4. The intelligent digital information acquisition system for precast tunnel segments at the factory end based on structured light according to claim 1, characterized in that: The X-axis sliding module I (2), Y-axis sliding module (3), Z-axis sliding module I (4), Z-axis sliding module II (5) and X-axis sliding module II (8) are all controlled by a closed-loop stepper motor for linear motion. The X-axis sliding module I (2) and X-axis sliding module II (8) include a base I and a slide I; the upper end face of the base I is provided with a pair of guide rails; the slide I is slidably mounted on the upper end face of the base I through the guide rails; the upper end face of the base I is also provided with a rack located between the pair of guide rails; the length direction of the rack is parallel to the length direction of the guide rails; the upper end face of the slide I is provided with a mounting through hole; a motor mounting flange plate is fixedly installed in the mounting through hole; a drive motor is provided on the motor mounting flange plate; a gear is provided below the motor mounting flange plate, and the drive motor is connected to the gear; the gear meshes with the rack. The Y-axis sliding module (3) includes a base II and a slide II; the side wall of the base II is provided with a pair of guide rails; the slide II is slidably mounted on the side wall of the base II via the guide rails; the side wall of the base II is also provided with a rack located between the pair of guide rails; the length direction of the rack is parallel to the length direction of the guide rails; the side wall of the slide II is provided with a mounting through hole; a motor mounting flange plate is fixedly installed in the mounting through hole; a drive motor is provided on the motor mounting flange plate; a gear is provided on the side of the motor mounting flange plate, and the drive motor is connected to the gear; the gear meshes with the rack. The Z-axis sliding module I (4) and Z-axis sliding module II (5) include a base III and a slide III; the base III is arranged vertically; the slide III is fixed on the slide II; a pair of guide rails are provided on the side wall of the base III near the slide II; the slide III is slidably installed on the side wall of the base III through the guide rails; the side wall of the base III is also provided with a rack located between the pair of guide rails; the length direction of the rack is parallel to the length direction of the guide rails; the side wall of the slide III has an installation through hole; a motor mounting flange plate is fixedly installed in the installation through hole; a drive motor is provided on the motor mounting flange plate; a gear is provided on the side of the motor mounting flange plate, and the drive motor is connected to the gear; the gear meshes with the rack.

5. The intelligent digital information acquisition system for precast tunnel segments at the factory end based on structured light according to claim 1, characterized in that: The support frame (1) is constructed by splicing horizontal and vertical aluminum alloy profiles into an open hexahedral structure; the inner wall of the support frame (1) integrates an infrared sensor array.

6. The acquisition method of the intelligent acquisition system for digital information of precast tunnel segments at the factory end based on structured light, as described in any one of claims 1 to 5, is characterized in that, Includes the following steps: 1) Place the precast tube segment (15) at the acquisition position; control the closed-loop stepper motor movement through PLC to move the sliding module XYZ three axes to the point cloud data acquisition points on both sides of the precast tube segment; 2) Adjust the horizontal and vertical angles of the two-dimensional rotating gimbal (6) through the host computer to ensure that the structured light cameras (7) on both sides of the precast tube segment accurately position the precast tube segment (15) to be scanned; 3) Use a structured light camera (7) to scan the precast tube segment (15) and obtain single-frame point cloud data; 4) Use statistical filtering algorithms to denoise the point cloud data and remove sparsely distributed outliers in the space; 5) Use a voxel downsampling algorithm to reduce the point cloud density and obtain downsampled single-frame point cloud data; 6) The translation data of each sliding module and the rotation angle of the two-dimensional gimbal are obtained through the three-axis movable rotatable motion control module. Then, the homogeneous coordinate transformation matrix between two adjacent frames of point clouds from the two structured light cameras is calculated to achieve coarse stitching between two adjacent point clouds. 7) Use the GICP algorithm to achieve accurate registration of two adjacent point clouds, and repeat the above steps to obtain the overall point cloud data on both sides of the precast segment. Finally, use the GICP algorithm again to accurately register the two point clouds to obtain the overall point cloud data of the surface of the precast segment.

7. The intelligent acquisition method for digital information of precast tunnel segments at the factory end based on structured light according to claim 6, characterized in that, Step 4) specifically includes the following sub-steps: 4.1) Calculate the average distance from each point in the input point cloud to its k nearest neighbors. Obtain a Gaussian distribution of the average distances of the k nearest neighbors by statistical analysis, and calculate the mean μ and standard deviation σ of this Gaussian distribution to determine the distance threshold thresh_d, where thresh_d = μ + k·σ (k = 1, 2, 3); 4.2) Based on the distance threshold determined in the first step, traverse each point in the input point cloud. If the average distance of the k neighboring points of a point is greater than the distance threshold thresh_d, then the point is identified as an outlier and deleted, thus achieving the noise reduction processing of the input point cloud.

8. The intelligent acquisition method for digital information of precast tunnel segments at the factory end based on structured light according to claim 6, characterized in that: In step 5), based on the three-dimensional coordinate set of the point cloud data, the maximum value x on the three coordinate axes (x, y, z) is calculated. max y max z max and minimum value x min y min z min Design the side length r of the voxel mesh, based on x max y max z max x min y min z min Calculate the side length l of the minimum bounding box of the point cloud. x l y l z : Calculate the dimensions of the voxel mesh; In the formula, Indicates rounding down; Calculate the index h of each point in the point cloud within the voxel grid; In the formula, h x h y h z represents the 3D index of the voxel mesh, and h represents the linear index of the voxel mesh; Store the linear index h of each point in an array in ascending order for later access; based on the three-dimensional index h of each point... x h y h z Each point is assigned to a corresponding voxel grid, the centroid of each voxel grid is calculated, and the centroid point replaces all points in the grid; finally, the point cloud data after voxel downsampling is obtained, realizing the lightweighting of point cloud data.

9. The intelligent acquisition method for digital information of precast tunnel segments at the factory end based on structured light according to claim 6, characterized in that: In step 6), the rotation matrix around the x-axis of the 3D structured light scanner coordinate system is calculated based on the pitch and rotation angle α of the 2D gimbal. Calculate the rotation matrix around the y-axis of the 3D structured light scanner based on the horizontal rotation angle β of the 2D gimbal. Using the above rotation matrix and The coordinates P of the second point cloud in the 3D structured light scanner coordinate system are... A Transform to the coordinates P of the first point cloud in the 3D structured light scanner coordinate system B ; By calculating the displacement changes t = (Δx, Δy, Δz) of the XYZ three-axis sliding module in the x, y, and z axes of the 3D structured light scanner coordinate system, and combining this with the aforementioned rotation matrix, an augmented matrix is ​​formed. Adding a row vector (0,0,0,1) to the augmented matrix yields the homogeneous coordinate transformation matrix between two adjacent point clouds, thus realizing the coordinate transformation of the two point clouds in the overall device coordinate system: By transforming the coordinates of two adjacent point clouds to the overall coordinate system specified by the scanning system using a homogeneous coordinate transformation matrix, coarse stitching between adjacent point clouds is achieved. The stitching effect is affected by the accuracy of the device.

10. The intelligent acquisition method for digital information of precast tunnel segments at the factory end based on structured light according to claim 6, characterized in that, Step 7) specifically includes the following sub-steps: 7.1) Given two sets of point cloud data, including source point cloud P and target point cloud Q, initialize the rigid transformation matrix T; 7.2) Calculate the covariance matrix. For each point p∈P and q∈Q, calculate the covariance matrices ΣP and ΣQ based on its neighborhood points: In the formula, μ P It is the centroid of a point within the domain of p; 7.3) Use the nearest neighbor algorithm to find the nearest neighbor pair (p,q) in the target point cloud Q for each point in the source point cloud P; 7.4) Calculate the error function based on the Gaussian distribution form of the two points: In the formula, (p,q) are point pairs; ΣP and ΣQ are the covariance matrices of each point pair, describing their local geometric distribution; (ΣP+ΣQ) -1 It is the inverse of the total covariance matrix, used to measure the error between point pairs. 7.5) Iteratively optimize the rigid transformation matrix T by minimizing the error function E: T=argmin T E 7.6) Apply the currently calculated rigid transformation matrix T to update the position of the source point cloud P, check the convergence condition of the error function, and if the error drops to a certain threshold or the number of iterations reaches a preset value, the accurate registration of the two frame point clouds is finally achieved.

Images