A device and method for fitting steel structure based on point cloud data

Abstract

The present invention relates to the technical field related to point cloud data fitting of steel structures, and specifically to a device and method for fitting steel structures based on point cloud data, comprising an unmanned aerial vehicle, a three-dimensional scanner, a dustproof structure and a self-cleaning structure. The unmanned aerial vehicle is provided with an installation groove on the lower surface, a mounting seat is fixedly installed in the installation groove, the three-dimensional scanner is installed on the mounting seat through a connecting piece, the dustproof structure is installed on the mounting seat, and the three-dimensional scanner is arranged in an inner cavity of the dustproof structure, and the self-cleaning structure is used to clean the outer surface of the dustproof structure. By setting a device for fitting steel structures based on point cloud data composed of a combination of an unmanned aerial vehicle, a three-dimensional scanner, a dustproof structure and a self-cleaning structure, dust adhesion can be avoided to affect the normal scanning of the three-dimensional scanner through the protection of the dustproof structure and the self-cleaning effect of the self-cleaning structure, so as to ensure the scanning accuracy of the three-dimensional scanner.

Description

Technical Field

[0001] The present invention relates to the technical field related to point cloud data fitting of steel structures, and in particular to a device and method for fitting steel structures based on point cloud data. Background Art

[0002] Point cloud is a set of data points in space. These points contain rich information, including three-dimensional coordinates X, Y, Z, color, classification value, intensity value, time, etc. Point cloud is a massive collection of points that express the spatial distribution of targets and the surface characteristics of targets in the same spatial reference system. Point cloud is generally obtained by 3D scanners, which measure many points on the external surface of objects and around these objects. Point cloud is widely used in many fields, including creating 3D CAD models for industrial production, for measurement and quality inspection, and for a large number of visual, animation, rendering and mass customization applications;

[0003] In the process of fitting the steel structure through point cloud data, it is necessary to scan and collect on-site data. However, since the construction site is usually a high-dust environment, and the scanning lens of the traditional scanning and acquisition structure is not provided with a special protective structure, it is easy for dust to adhere to the lens during the scanning process, thereby affecting the data acquisition accuracy of the scanning and acquisition device, thereby having a certain impact on subsequent projects. For this reason, the present invention proposes a device and method for fitting steel structure based on point cloud data to solve the above problem. Summary of the invention

[0004] The purpose of the present invention is to provide a device and method for fitting steel structures based on point cloud data to solve the problems raised in the above background technology.

[0005] To achieve the above object, the present invention provides the following technical solution: a device for fitting a steel structure based on point cloud data, comprising:

[0006] A drone, wherein a mounting groove is provided on the lower surface of the drone, and a mounting seat is fixedly installed in the mounting groove;

[0007] A three-dimensional scanner, wherein the three-dimensional scanner is mounted on a mounting base via a connecting piece;

[0008] A dustproof structure, wherein the dustproof structure is mounted on a mounting seat, and the three-dimensional scanner is disposed in an inner cavity of the dustproof structure;

[0009] The self-cleaning structure is used to clean the outer surface of the dustproof structure.

[0010] Preferably, the mounting seat is arranged in a square tube structure, the dustproof structure is a transparent glass ball with openings at both ends, and a rotating tube is integrally formed at the openings at both ends of the dustproof structure, and the rotating tube is rotatably mounted on the mounting seat through a ball bearing.

[0011] Preferably, the connecting member is composed of a connecting shaft, a connecting seat and a connecting shaft positioning seat. The connecting shaft is positioned on the mounting seat through the connecting shaft positioning seat. The connecting seat is integrally formed with the end of the connecting shaft. The three-dimensional scanner is fixedly mounted on the connecting seat, and the wire body on the three-dimensional scanner passes through the wiring hole on the connecting shaft. The connecting shaft is rotatably connected to the rotating tube on one side through a sealed bearing.

[0012] Preferably, a ring gear is integrally formed at the end of the rotating tube, a servo motor is fixedly mounted on the mounting seat, a transmission gear is fixedly mounted on the output shaft of the servo motor, the transmission gear is meshed with the ring gear, and the servo motor is connected to the control module on the drone via electrical signals.

[0013] Preferably, the self-cleaning structure includes a cleaning cloth, an absorbent cotton felt, a cleaning liquid storage box and a cleaning nozzle combination, the cleaning cloth, the absorbent cotton felt and the outer side surface of the dustproof structure are arranged in contact with each other, the mounting seat is composed of a first seat body and a second seat body, the side edges of the first seat body and the second seat body are respectively integrally formed with a primary connecting ear plate and a secondary connecting ear plate, the primary connecting ear plate and the secondary connecting ear plate are positioned and connected by bolts and nuts, and a primary screw seat and a secondary screw seat are respectively integrally formed on the inner side walls of the first seat body and the second seat body.

[0014] Preferably, the first-level screw seat is connected to a first-level positioning seat via a screw structure, and the second-level screw seat is connected to a second-level positioning seat via a screw structure. The cleaning cloth and the absorbent cotton felt are fixedly connected to the inner walls of the first-level positioning seat and the second-level positioning seat respectively. The cleaning liquid storage box is fixed on the upper end of the first seat body, and the cleaning nozzle is fixed on the lower surface of the cleaning liquid storage box. The liquid inlet of the cleaning nozzle is connected to the liquid pump inside the cleaning liquid storage box through an infusion tube, and the liquid pump is controlled by a control module on the drone.

[0015] Preferably, an alignment groove is provided on the side of the primary connecting ear plate, and an alignment seat is integrally formed on the side of the secondary connecting ear plate. When the primary connecting ear plate and the secondary connecting ear plate are actually docked, the alignment seat is embedded in the alignment groove.

[0016] A method for fitting a steel structure based on point cloud data, wherein the method for fitting a steel structure based on point cloud data is implemented by the above-mentioned device for fitting a steel structure based on point cloud data, and the method comprises:

[0017] Step 1: Scan and collect on-site data;

[0018] Step 2: Process point cloud data;

[0019] Step 3: Establish theoretical entity model of steel structure;

[0020] Step 4: The point cloud data is segmented according to a single steel beam;

[0021] Step 5: Clear discrete points;

[0022] Step 6: Match point clouds using annealing algorithm;

[0023] Step 7: Return to 3D space for inspection.

[0024] Preferably, the processing of point cloud data comprises the following steps:

[0025] a. Point cloud registration: Automatically register multiple collected point cloud data by placing target points during scanning. For stations without target points, use the common point finding and feature matching methods to perform registration and merge all station clouds into a whole.

[0026] b. Point cloud denoising: Use distance threshold or statistical algorithms to denoise point cloud data and remove discrete points that do not match the subject;

[0027] c. Point cloud segmentation: Delete irrelevant point clouds around the target area and extract target objects in the point cloud data, such as steel structures and structures;

[0028] d. Export point cloud: export the point cloud to pts format and import it into Rhino software for further operation;

[0029] The method of establishing a theoretical entity model of a steel structure comprises the following steps:

[0030] e. Extract the centerline of the steel structure: including extracting the centerline of the steel structure from the Tekla model of the steel structure, importing the centerline of the steel structure into the Rhino software; establishing the centerline model of the steel structure in the Rhino software according to the steel structure drawings;

[0031] f. Establish a three-dimensional model of the steel structure: In Rhino software, according to the steel structure specifications on the drawings, establish a theoretical model of the steel structure based on the center line of the steel structure.

[0032] Preferably, the point cloud data is segmented according to a single steel beam, comprising the following steps:

[0033] g. Create a bounding box for a single steel beam: In Rhino software, split the steel structure solid model in step 3 into segments, use the grasshopper plug-in, and use the "Bounding Box" component to create a bounding box;

[0034] h. Segment the point cloud: Use the Grasshopper plug-in and the "Cloud Box Crop" component to segment the point cloud into individual steel beams;

[0035] The clearing of discrete points comprises:

[0036] j. Clear discrete points: Use the "CloudSOR" component to clear discrete points and points that do not belong to the target steel structure to obtain an accurate point group;

[0037] k. Convert point cloud to points: Use the "Cloud Explode" component to explode the point cloud into individual points;

[0038] The method of matching point clouds using an annealing algorithm includes:

[0039] l. Establish a conversion coordinate system: Take the center of the steel structure solid model as the origin, extract the longer side of the steel structure as the X-axis, and the shorter side as the Y-axis, and establish a three-dimensional conversion coordinate system;

[0040] m. Establish annealing algorithm process: Use the "Galapagos Evolutionary Solver" annealing algorithm component, set five decimal parameters as input parameters, respectively control the X-axis, Y-axis, and Z-axis movement distances of the steel structure entity model in the three-dimensional transformation coordinate system and the XY plane and XZ plane rotation angles, and use the sum of the distances from the point group obtained in step 5 to the steel structure entity model as the control parameter;

[0041] n. Convert the steel structure entity model: Set the "Galapagos Evolutionary Solver" annealing algorithm component, connect the input parameters, set the calculation result to approach the minimum value of the control parameter, set the maximum operation time, and perform data operations. The annealing algorithm component will perform a series of random parameter changes. In each iteration step, the algorithm will try to transfer from the current solution to a new solution based on a certain probability function. If the new solution is better than the old solution (that is, the objective function value is lower), the new solution is directly accepted; if the new solution is worse, the worse solution is accepted with a certain probability, and the minimum value of the control parameter is continuously approached. At the end of the settlement, the input parameters when the control parameter is at the minimum value will be taken to obtain the virtual model of the steel structure entity conversion completed;

[0042] The returning to the 3D space for inspection includes:

[0043] o. Bake into Rhino software: Bake the virtual model of the steel structure entity in Grasshopper into Rhino software to build the entity for inspection;

[0044] p. Convert sequentially: Use the loop control component to convert the steel structure solid model sequentially according to the segmentation in step 4 and bake it into the Rhino software.

[0045] Compared with the prior art, the present invention has the following beneficial effects:

[0046] 1. By setting up a device for fitting steel structure based on point cloud data, which is composed of a drone, a 3D scanner, a dustproof structure and a self-cleaning structure, the dust adhesion is prevented from affecting the normal scanning of the 3D scanner through the protection of the dustproof structure and the self-cleaning effect of the self-cleaning structure, so as to ensure the scanning accuracy of the 3D scanner;

[0047] 2. By setting the self-cleaning structure to be composed of a cleaning cloth, absorbent cotton felt, a cleaning liquid storage box and a cleaning nozzle, the self-cleaning effect is improved through the combined action of the cleaning cloth, absorbent cotton felt, a cleaning liquid storage box and a cleaning nozzle, thereby further improving the scanning accuracy of the three-dimensional scanner.

[0048] 3. With the help of 3D scanners, it is possible to efficiently and accurately collect wide-area point cloud data, covering the full range of 3D features of the main structure and steel structure on site. The laser radar measurement technology used not only ensures high accuracy and high density of data, but also can accurately capture complex details such as curved arcs or diameter changes. Compared with the traditional method that relies on manual measurement or aerial photography, this solution can provide a more detailed and accurate information set. Combined with laser radar technology, it can perform rapid and automated data collection tasks for the entire building, greatly shortening the data collection cycle, thereby accelerating the progress of the project and improving work efficiency;

[0049] 4. Due to the uncertainty in on-site scanning, such as the front and back occlusion of steel structures and temporary components, it is impossible to scan all the front and back sides of the steel structure, and some of the point clouds are missing or omitted. The present invention uses the annealing algorithm to obtain the minimum sum of the point cloud distance entities, which can avoid interference caused by missing point clouds or the presence of clutter in the point clouds, try multiple value ranges, and through a series of random state changes, ensure that the output inverse model is the global optimal solution for the existing point clouds;

[0050] 5. The present invention can be applied to H-shaped, polygonal and other special-shaped curved steel structures, and can also be applied to similar shaped structures, such as water supply and drainage pipes, and electromechanical installation bridges. The reverse modeling method of the present invention can also be used to provide a more detailed and accurate model for subsequent professional modeling such as curtain walls. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 It is a schematic diagram of the upper structure of the present invention;

[0052] Figure 2 It is a schematic diagram of the lower structure of the present invention;

[0053] Figure 3 This is a schematic diagram of the mounting seat structure of the present invention;

[0054] Figure 4 This is a schematic diagram of the first seat structure of the present invention;

[0055] Figure 5It is a schematic diagram of the second seat structure of the present invention;

[0056] Figure 6 This is a schematic diagram of the self-cleaning structure distribution of the present invention;

[0057] Figure 7 The figure is a flow chart of the method of the present invention.

[0058] In the figure: drone 1, mounting seat 2, 3D scanner 3, dustproof structure 4, first seat body 5, second seat body 6, primary screw seat 7, primary positioning seat 8, cleaning cloth 9, secondary positioning seat 10, absorbent cotton felt 11, cleaning liquid storage box 12, cleaning nozzle 13, rotating tube 14, ball bearing 15, ring gear 16, sealed bearing 17, connecting shaft 18, connecting seat 19, connecting shaft positioning seat 20, line body 21, primary connecting ear plate 22, secondary connecting ear plate 23, alignment groove 24, alignment seat 25. DETAILED DESCRIPTION

[0059] In order to make the purpose and technical solution of the present invention clearly and completely described, and the advantages more clearly understood, the embodiments of the present invention are further described in detail with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are part of the embodiments of the present invention, rather than all of the embodiments, and are only used to explain the embodiments of the present invention, and are not used to limit the embodiments of the present invention. All other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0060] See also Figure 1-7 The present invention provides the following three preferred embodiments:

[0061] Embodiment 1, a device for fitting a steel structure based on point cloud data, comprises a drone 1, a three-dimensional scanner 3, a dustproof structure 4 and a self-cleaning structure, a mounting groove is provided on the lower surface of the drone 1, a mounting seat 2 is fixedly installed in the mounting groove, the three-dimensional scanner 3 is installed on the mounting seat 2 through a connecting piece, the dustproof structure 4 is installed on the mounting seat 2, and the three-dimensional scanner 3 is arranged in the inner cavity of the dustproof structure 4, and the self-cleaning structure is used to clean the outer surface of the dustproof structure 4.

[0062] The mounting seat 2 is arranged in a square tube structure, the dustproof structure 4 is a transparent glass ball with openings at both ends, and a rotating tube 14 is integrally formed at the openings at both ends of the dustproof structure 4 , and the rotating tube 14 is rotatably mounted on the mounting seat 2 through a ball bearing 15 .

[0063] The connecting part is composed of a connecting shaft 18, a connecting seat 19 and a connecting shaft positioning seat 20. The connecting shaft 18 is positioned on the mounting seat 2 through the connecting shaft positioning seat 20. The connecting seat 19 and the end of the connecting shaft 18 are integrally formed. The three-dimensional scanner 3 is fixedly installed on the connecting seat 19, and the wire body 21 on the three-dimensional scanner 3 is passed through the wiring hole on the connecting shaft 18. The connecting shaft 18 is rotatably connected to the rotating tube 14 on one side through a sealed bearing 17.

[0064] A ring gear 16 is integrally formed at the end of the rotating tube 14, a servo motor is fixedly mounted on the mounting seat 2, a transmission gear is fixedly mounted on the output shaft of the servo motor, the transmission gear is meshed with the ring gear 16, the servo motor is electrically connected to the control module on the drone 1, and a device for fitting a steel structure based on point cloud data is arranged, which is composed of the drone 1, the three-dimensional scanner 3, the dustproof structure 4 and the self-cleaning structure, so that dust adhesion can be avoided to affect the normal scanning of the three-dimensional scanner 3 through the protection of the dustproof structure 4 and the self-cleaning effect of the self-cleaning structure, so as to ensure the scanning accuracy of the three-dimensional scanner 3.

[0065] Embodiment 2, on the basis of embodiment 1, the self-cleaning structure includes a cleaning cloth 9, an absorbent cotton felt 11, a cleaning liquid storage box 12 and a cleaning nozzle 13, the cleaning cloth 9, the absorbent cotton felt 11 are arranged in contact with the outer side of the dustproof structure 4, the mounting seat 2 is composed of a first seat body 5 and a second seat body 6, the side edges of the first seat body 5 and the second seat body 6 are respectively integrally formed with a primary connecting ear plate 22 and a secondary connecting ear plate 23, the primary connecting ear plate 22 and the secondary connecting ear plate 23 are positioned and connected by bolts and nuts, and the inner side walls of the first seat body 5 and the second seat body 6 are respectively integrally formed with a primary screw seat 7 and a secondary screw seat.

[0066] A first-level screw seat 7 is connected to a first-level positioning seat 8 through a screw structure, and a second-level screw seat is connected to a second-level positioning seat 10 through a screw structure. A cleaning cloth 9 and an absorbent cotton felt 11 are fixedly connected to the inner side walls of the first-level positioning seat 8 and the second-level positioning seat 10 respectively. A cleaning liquid storage box 12 is fixed to the upper side end of the first seat body 5, and a cleaning nozzle 13 is fixed to the lower surface of the cleaning liquid storage box 12. The liquid inlet of the cleaning nozzle 13 is connected to the liquid pump inside the cleaning liquid storage box 12 through an infusion tube. The liquid pump is controlled by a control module on the drone 1. By setting the self-cleaning structure to be composed of a cleaning cloth 9, an absorbent cotton felt 11, a cleaning liquid storage box 12 and a cleaning nozzle 13, the self-cleaning effect is improved through the combined action of the cleaning cloth 9, the absorbent cotton felt 11, the cleaning liquid storage box 12 and the cleaning nozzle 13, thereby further improving the scanning accuracy of the three-dimensional scanner 3.

[0067] An alignment groove 24 is provided on the side of the primary connecting ear plate 22, and an alignment seat 25 is integrally formed on the side of the secondary connecting ear plate 23. When the primary connecting ear plate 22 and the secondary connecting ear plate 23 are actually docked, the alignment seat 25 is embedded in the alignment groove 24. The setting of the alignment seat 25 and the alignment groove 24 can improve the connection stability of the primary connecting ear plate 22 and the secondary connecting ear plate 23.

[0068] Embodiment 3, based on Embodiment 2, a method for fitting a steel structure based on point cloud data is provided. The method for fitting a steel structure based on point cloud data is implemented by the above-mentioned device for fitting a steel structure based on point cloud data, and the method includes:

[0069] Step 1: Scan and collect on-site data;

[0070] Step 2: Process point cloud data;

[0071] Step 3: Establish theoretical entity model of steel structure;

[0072] Step 4: The point cloud data is segmented according to a single steel beam;

[0073] Step 5: Clear discrete points;

[0074] Step 6: Match point clouds using annealing algorithm;

[0075] Step 7: Return to 3D space for inspection;

[0076] The 3D scanner can efficiently and accurately collect wide-area point cloud data, covering the full range of 3D features of the main structure and steel structure on site. The laser radar measurement technology used not only ensures high accuracy and density of the data, but also can accurately capture complex details such as curved arcs or diameter changes. Compared with the traditional method that relies on manual measurement or aerial photography, this solution can provide a more detailed and accurate information set. Combined with laser radar technology, it can perform rapid and automated data collection tasks for the entire building, greatly shortening the data collection cycle, thereby accelerating the progress of the project and improving work efficiency.

[0077] Processing point cloud data includes the following steps:

[0078] a. Point cloud registration: Automatically register multiple collected point cloud data by placing target points during scanning. For stations without target points, use the common point finding and feature matching methods to perform registration and merge all station clouds into a whole.

[0079] b. Point cloud denoising: Use distance threshold or statistical algorithms to denoise point cloud data and remove discrete points that do not match the subject;

[0080] c. Point cloud segmentation: Delete irrelevant point clouds around the target area and extract target objects in the point cloud data, such as steel structures and structures;

[0081] d. Export point cloud: export the point cloud to pts format and import it into Rhino software for further operation;

[0082] Establishing a theoretical entity model of a steel structure includes the following steps:

[0083] e. Extract the centerline of the steel structure: including extracting the centerline of the steel structure from the Tekla model of the steel structure, importing the centerline of the steel structure into the Rhino software; establishing the centerline model of the steel structure in the Rhino software according to the steel structure drawings;

[0084] f. Establish a three-dimensional model of the steel structure: In Rhino software, according to the steel structure specifications on the drawings, establish a theoretical model of the steel structure based on the center line of the steel structure.

[0085] The point cloud data is segmented according to a single steel beam, including the following steps:

[0086] g. Create a bounding box for a single steel beam: In Rhino software, split the steel structure solid model in step 3 into segments, use the grasshopper plug-in, and use the "Bounding Box" component to create a bounding box;

[0087] h. Segment the point cloud: Use the Grasshopper plug-in and the "Cloud Box Crop" component to segment the point cloud into individual steel beams;

[0088] Clear discrete points, including:

[0089] j. Clear discrete points: Use the "CloudSOR" component to clear discrete points and points that do not belong to the target steel structure to obtain an accurate point group;

[0090] k. Convert point cloud to points: Use the "Cloud Explode" component to explode the point cloud into individual points;

[0091] Match point clouds using an annealing algorithm, including:

[0092] l. Establish a conversion coordinate system: Take the center of the steel structure solid model as the origin, extract the longer side of the steel structure as the X-axis, and the shorter side as the Y-axis, and establish a three-dimensional conversion coordinate system;

[0093] m. Establish annealing algorithm process: Use the "Galapagos Evolutionary Solver" annealing algorithm component, set five decimal parameters as input parameters, respectively control the X-axis, Y-axis, and Z-axis movement distances of the steel structure entity model in the three-dimensional transformation coordinate system and the XY plane and XZ plane rotation angles, and use the sum of the distances from the point group obtained in step 5 to the steel structure entity model as the control parameter;

[0094] n. Convert the steel structure entity model: Set the "Galapagos Evolutionary Solver" annealing algorithm component, connect the input parameters, set the calculation result to approach the minimum value of the control parameter, set the maximum operation time, and perform data operations. The annealing algorithm component will perform a series of random parameter changes. In each iteration step, the algorithm will try to transfer from the current solution to a new solution based on a certain probability function. If the new solution is better than the old solution (that is, the objective function value is lower), the new solution is directly accepted; if the new solution is worse, the worse solution is accepted with a certain probability, and the minimum value of the control parameter is continuously approached. At the end of the settlement, the input parameters when the control parameter is at the minimum value will be taken to obtain the virtual model of the steel structure entity conversion completed;

[0095] Return to 3D space for inspection, including:

[0096] o. Bake into Rhino software: Bake the virtual model of the steel structure entity in Grasshopper into Rhino software to build the entity for inspection;

[0097] p. Convert sequentially: Use the loop control component to convert the steel structure solid model sequentially according to the segmentation in step 4 and bake it into the Rhino software.

[0098] Although the above describes the illustrative specific implementation methods of the present application so that technicians in this technical field can understand the present application, the present application is not limited to the scope of the specific implementation methods. For ordinary technicians in this technical field, as long as various changes are within the spirit and scope of the present application defined and determined by the attached claims, all application creations using the concept of the present application are protected.

Claims

1. A device for fitting steel structure based on point cloud data, characterized in that: include: A drone (1), wherein a mounting groove is provided on a lower surface of the drone (1), and a mounting seat (2) is fixedly mounted in the mounting groove; A three-dimensional scanner (3), wherein the three-dimensional scanner (3) is mounted on the mounting base (2) via a connecting piece; A dustproof structure (4), wherein the dustproof structure (4) is mounted on the mounting seat (2), and the three-dimensional scanner (3) is disposed in an inner cavity of the dustproof structure (4); A self-cleaning structure, wherein the self-cleaning structure is used to clean the outer surface of the dustproof structure (4).

2. The device for fitting steel structure based on point cloud data according to claim 1, characterized in that: The mounting seat (2) is arranged in a square tube structure, the dustproof structure (4) is a transparent glass ball with openings at both ends, and a rotating tube (14) is integrally formed at the openings at both ends of the dustproof structure (4), and the rotating tube (14) is rotatably mounted on the mounting seat (2) via a ball bearing (15).

3. The device for fitting steel structure based on point cloud data according to claim 2, characterized in that: The connecting member is composed of a connecting shaft (18), a connecting seat (19) and a connecting shaft positioning seat (20); the connecting shaft (18) is positioned on the mounting seat (2) through the connecting shaft positioning seat (20); the connecting seat (19) and the end of the connecting shaft (18) are integrally formed; the three-dimensional scanner (3) is fixedly mounted on the connecting seat (19); and the wire body (21) on the three-dimensional scanner (3) passes through a wire routing hole on the connecting shaft (18); and the connecting shaft (18) is rotatably connected to a rotating tube (14) on one side through a sealing bearing (17).

4. The device for fitting steel structure based on point cloud data according to claim 3, characterized in that: A ring gear (16) is integrally formed at the end of the rotating tube (14); a servo motor is fixedly mounted on the mounting seat (2); a transmission gear is fixedly mounted on the output shaft of the servo motor; the transmission gear is meshed with the ring gear (16); and the servo motor is electrically connected to a control module on the drone (1).

5. The device for fitting steel structure based on point cloud data according to claim 4, characterized in that: The self-cleaning structure comprises a cleaning cloth (9), a water-absorbent cotton felt (11), a cleaning liquid storage box (12) and a cleaning nozzle (13) in combination. The cleaning cloth (9) and the water-absorbent cotton felt (11) are arranged in contact with the outer side surface of the dustproof structure (4). The mounting seat (2) is composed of a first seat body (5) and a second seat body (6). The first seat body (5) and the second seat body (6) are respectively integrally formed with a primary connecting ear plate (22) and a secondary connecting ear plate (23) on their side edges. The primary connecting ear plate (22) and the secondary connecting ear plate (23) are positioned and connected by bolts and nuts. The first seat body (5) and the second seat body (6) are respectively integrally formed with a primary screw seat (7) and a secondary screw seat on their inner side walls.

6. The device for fitting steel structure based on point cloud data according to claim 5, characterized in that: The primary screw seat (7) is connected to a primary positioning seat (8) via a screw structure, and the secondary screw seat is connected to a secondary positioning seat (10) via a screw structure. The cleaning cloth (9) and the absorbent cotton felt (11) are fixedly connected to the inner side walls of the primary positioning seat (8) and the secondary positioning seat (10), respectively. The cleaning liquid storage box (12) is fixed to the upper side end of the first seat body (5), and the cleaning nozzle (13) is fixed to the lower surface of the cleaning liquid storage box (12). The liquid inlet of the cleaning nozzle (13) is connected to the liquid pump inside the cleaning liquid storage box (12) via an infusion tube, and the liquid pump is controlled by a control module on the drone (1).

7. The device for fitting steel structure based on point cloud data according to claim 6, characterized in that: The side surface of the primary connecting ear plate (22) is provided with an alignment groove (24), and the side surface of the secondary connecting ear plate (23) is integrally formed with an alignment seat (25); when the primary connecting ear plate (22) and the secondary connecting ear plate (23) are actually connected, the alignment seat (25) is embedded in the alignment groove (24).

8. A method for fitting a steel structure based on point cloud data, characterized in that: The method for fitting a steel structure based on point cloud data is implemented by any one of the devices for fitting a steel structure based on point cloud data according to claims 1 to 7, and the method comprises: Step 1: Scan and collect on-site data; Step 2: Process point cloud data; Step 3: Establish theoretical entity model of steel structure; Step 4: Divide the point cloud data into sections according to a single steel beam; Step 5: Clear discrete points; Step 6: Match point clouds using annealing algorithm; Step 7: Return to 3D space for inspection.

9. The method for fitting a steel structure based on point cloud data according to claim 8, characterized in that: The processing of point cloud data comprises the following steps: a. Point cloud registration: Automatically register multiple collected point cloud data by placing target points during scanning. For stations without target points, use the common point finding and feature matching methods to perform registration and merge all station clouds into a whole. b. Point cloud denoising: Use distance threshold or statistical algorithms to denoise point cloud data and remove discrete points that do not match the subject; c. Point cloud segmentation: Delete irrelevant point clouds around the target area and extract target objects in the point cloud data, such as steel structures and structures; d. Export point cloud: export the point cloud to pts format and import it into Rhino software for further operation; The method of establishing a theoretical entity model of a steel structure comprises the following steps: e. Extract the centerline of the steel structure: including extracting the centerline of the steel structure from the Tekla model of the steel structure, importing the centerline of the steel structure into the Rhino software; establishing the centerline model of the steel structure in the Rhino software according to the steel structure drawings; f. Establish a three-dimensional model of the steel structure: In Rhino software, according to the steel structure specifications on the drawings, establish a theoretical model of the steel structure based on the center line of the steel structure.

10. The method for fitting a steel structure based on point cloud data according to claim 9, characterized in that: The point cloud data is segmented according to a single steel beam, including the following steps: g. Create a bounding box for a single steel beam: In Rhino software, split the steel structure solid model in step 3 into segments, use the grasshopper plug-in, and use the "Bounding Box" component to create a bounding box; h. Segment the point cloud: Use the Grasshopper plug-in and the "Cloud Box Crop" component to segment the point cloud into individual steel beams; The clearing of discrete points comprises: j. Clear discrete points: Use the "CloudSOR" component to clear discrete points and points that do not belong to the target steel structure to obtain an accurate point group; k. Convert point cloud to points: Use the "Cloud Explode" component to explode the point cloud into individual points; The method of matching point clouds using an annealing algorithm includes: l. Establish a conversion coordinate system: Take the center of the steel structure solid model as the origin, extract the longer side of the steel structure as the X-axis, and the shorter side as the Y-axis, and establish a three-dimensional conversion coordinate system; m. Establish annealing algorithm process: Use the "Galapagos Evolutionary Solver" annealing algorithm component, set five decimal parameters as input parameters, respectively control the X-axis, Y-axis, and Z-axis movement distances of the steel structure entity model in the three-dimensional transformation coordinate system and the XY plane and XZ plane rotation angles, and use the sum of the distances from the point group obtained in step 5 to the steel structure entity model as the control parameter; n. Convert the steel structure entity model: Set the "Galapagos Evolutionary Solver" annealing algorithm component, connect the input parameters, set the calculation result to approach the minimum value of the control parameter, set the maximum operation time, and perform data operations. The annealing algorithm component will perform a series of random parameter changes. In each iteration step, the algorithm will try to transfer from the current solution to a new solution based on a certain probability function. If the new solution is better than the old solution (that is, the objective function value is lower), the new solution is directly accepted; if the new solution is worse, the worse solution is accepted with a certain probability, and the minimum value of the control parameter is continuously approached. At the end of the settlement, the input parameters when the control parameter is at the minimum value will be taken to obtain the virtual model of the steel structure entity conversion completed; The returning to the 3D space for inspection includes: o. Bake into Rhino software: Bake the virtual model of the steel structure entity in Grasshopper into Rhino software to build the entity for inspection; p. Convert sequentially: Use the loop control component to convert the steel structure solid model sequentially according to the segmentation in step 4 and bake it into the Rhino software.

Images