Digital detection system and method for skin attribute of large-size metal wallboard

Abstract

The invention provides a digitalized detection system and method for the skin attribute of a large-size metal wallboard. The system is characterized in that the system comprises a binocular optical tracker (1), a fixing tool (2), a two-dimensional industrial camera (3), a three-dimensional spherical scanner (4), a multi-axis mechanical arm (5), an automatic guided vehicle (6), a software analysis work station (7) and a safety net (8), the binocular optical tracker (1) achieves full coverage of a measurement tracking view field of the large-size metal wall plate, and the fixing tool (2) is used for fixing the large-size metal wall plate. The two-dimensional industrial camera (3) and the three-dimensional spherical scanner (4) are fixed to the tail axis of the multi-axis mechanical arm (5), and the software analysis work station (7) receives collected two-dimensional data and three-dimensional data and completes analysis work. According to the invention, the difficulties of large size of a wallboard workpiece, long manual detection time, large detection error and the like can be overcome, and the digitalized detection of the skin attribute of the large-size metal wallboard is realized based on a digitalized detection technology.

Description

Technical Field

[0001] The invention relates to the field of aviation detection technology, and in particular to a digital detection system and method for large-size metal wall panel skin properties. Background Art

[0002] With the development of aerospace technology, aerospace products have put forward higher requirements on metal wall panel skins.

[0003] In aerospace products, metal panel skins are usually made of lightweight and high-strength materials, and the base materials include aluminum, aluminum alloys, titanium and titanium alloys, copper, copper alloys or austenitic stainless steel. During design, in order to ensure the strength of the skin parts, one side is generally designed to be a smooth surface and the other side is a mesh surface with reinforcing ribs, so as to achieve weight reduction while meeting the strength requirements. The manufacture of metal panel skins involves forming, machining and milling meshes, etc. These steps require precise process control and quality inspection to ensure that the quality and performance of the metal panel skins meet the design requirements.

[0004] The manual inspection method currently used in the production process has low coverage of measurement indicators, low measurement efficiency, poor data stability, and low data informatization level, which is prone to missed inspections and poor inspection consistency. It is no longer suitable for the development trend of intelligent transformation and digital transformation engineering and the urgent need for high-quality production. Especially for large-sized aerospace product workpieces, the manual measurement process takes a long time and has large operating errors. The accuracy and efficiency of the measurement process need to be improved urgently. Summary of the invention

[0005] The purpose of this invention is to address the defects of the above-mentioned prior arts and propose a digital detection technology solution for large-size metal wall panel skins, so as to solve the problems of long detection process time, low coverage of detection indicators, and low level of overall process informatization of traditional methods, so as to meet the urgent needs of high-quality production of aerospace products and the development trend of intelligent transformation and digital transformation projects.

[0006] Specifically, the present invention provides a large-size metal wall panel skin property digital detection system, characterized in that it includes a binocular optical tracker 1, a fixed tool 2, a two-dimensional industrial camera 3, a three-dimensional spherical scanner 4, a multi-axis robotic arm 5, an automatic guided vehicle 6, a software analysis workstation 7 and a safety net 8. The binocular optical tracker 1 achieves full coverage of the measurement and tracking field of view of the large-size metal wall panel, and is used to determine the coordinates of the three-dimensional spherical scanner 4. The fixed tool 2 is placed at a position set within the field of view determined by the binocular optical tracker 1, and is used to place the wall panel to be inspected. The two-dimensional industrial camera 3 and the three-dimensional spherical scanner 4 are fixed on the end axis of the multi-axis robotic arm 5, and are used to measure the wall panel to be inspected. The automatic guided vehicle 6 and the multi-axis robotic arm 5 are assembled into a composite robot as a measurement bearing platform. The software analysis workstation 7 receives the collected two-dimensional data and three-dimensional data and completes the analysis work. The safety net 8 is used to enclose the inspection station to realize closed space management.

[0007] Furthermore, the binocular optical trackers 1 are arranged in three groups, and the three groups are respectively arranged at the center position of the inner side of the arc of the wall plate to be inspected and at the two ends of the outer side of the arc of the wall plate to be inspected.

[0008] Furthermore, the fixing fixture 2 is in an arc shape matching the wall panel to be inspected, and has a column at each end for clamping the wall panel to be inspected.

[0009] Furthermore, the multi-axis robot arm 5 has multiple axes and can adjust the end position in three-dimensional directions.

[0010] The present invention also provides a method for digitally detecting the properties of large-size metal wall panels, which uses the digital detection system for the properties of large-size metal wall panels according to claim 1, and is characterized in that it includes the following steps: S1, building a full-coverage measurement and tracking field of view based on a binocular optical tracker 1, and installing the wall panel to be measured on a fixed tooling 2 to achieve full field coverage of the wall panel to be measured; S2, assembling an automatic guided vehicle 6 and a multi-axis robotic arm 5 into a composite robot, and installing a three-dimensional spherical scanner 4 and a two-dimensional industrial camera 3 on the multi-axis robotic arm 5; S3, starting the composite robot, and according to the planned route, using the three-dimensional spherical scanner 4 to obtain three-dimensional point cloud data of the wall panel surface, and using the two-dimensional industrial camera 3 to obtain two-dimensional image data of the wall panel surface; S4, transmitting the obtained data to the software analysis workstation 7 to complete skin property analysis and defect detection; S5, generating a detection result attribute report, and evacuating the wall panel to be measured.

[0011] Furthermore, step S1 comprises the following steps:

[0012] Step S101: perform scanning field of view analysis and calculation according to the maximum size of the wall panel to be measured, equipment attributes and on-site environment; reasonably set the system footprint according to the maximum coverage of the binocular optical tracker 1, the motion range of the automatic guided vehicle 6 and the multi-axis mechanical arm 5, the effective distance of the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3; perform detailed calculation of the field of view coverage, and achieve full coverage of the measurement and tracking field of view of the wall panel to be measured;

[0013] Step S102: according to the analysis results, the fixed tooling 2 is constructed, and the position of the binocular optical tracker 1 is set to complete the construction of the tracking field of view, so as to realize the measurement and tracking field of view required for digital detection.

[0014] Furthermore, step S2 comprises the following steps:

[0015] Step S201: assemble the multi-axis robot arm 5 and the automatic guided vehicle 6 to build a measurement bearing platform, wherein the weight of the equipment to be carried by the automatic guided vehicle 6 and the height of the multi-axis robot arm 5 are analyzed to complete the construction of the bearing platform;

[0016] Step S202: Use a support fixture on the multi-axis robotic arm 5, and install a three-dimensional spherical scanner 4 and a two-dimensional industrial camera 3 on the two arms respectively.

[0017] Furthermore, step S3 comprises the following steps:

[0018] Step S301, according to the model of the wall panel to be tested, the automatic guided vehicle 6 adopts laser navigation, manually and reasonably plans the driving route and the movement of the multi-axis robot 5, and generates the path points of the movement of the multi-axis robot 5 according to the path planning results, each of which contains the position and posture information of the multi-axis robot 5, as well as the position and orientation information of the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3;

[0019] Step S302: Start the composite robot and execute it twice according to the three-dimensional scanning action route and the two-dimensional shooting action route respectively to obtain three-dimensional point cloud data and two-dimensional image data.

[0020] Furthermore, step S4 comprises the following steps:

[0021] Step S401, the three-dimensional point cloud data and the two-dimensional image data are transmitted to the workstation via a wireless network;

[0022] Step S402, using a three-dimensional point cloud algorithm to calculate the properties of the wall thickness and rib width based on the three-dimensional point cloud data;

[0023] Step S403: Use a deep learning algorithm to detect scratches and thread defects on the wall panel surface based on the two-dimensional image data.

[0024] Furthermore, in step S301: when planning the driving path, possible obstacles around the multi-axis robot 5, the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3 are considered, and the path planning is ensured to avoid obstacles by setting a suitable safety distance or performing collision detection; firstly, the starting point and the end point of the multi-axis robot 5, the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3 are determined to ensure that the entire wall panel surface is covered; the wall panel surface is divided into multiple small areas for more detailed scanning, and for each small area, a scanning path is defined to cover all points in the area, and the definition of the scanning path adopts the following strategy:

[0025] When the surface of the wall panel to be tested is a smooth surface, a serpentine scanning path is used for the area inside the smooth surface: starting from the starting point, scanning along the horizontal direction, and then when reaching the boundary, moving to the next line and changing the scanning direction, and so on, until reaching the end point.

[0026] The trajectory operation parameters of the three-dimensional spherical scanner 4 are: the distance from the area to be measured is within the optimal scanning distance of 300±30mm, the binoculars of the three-dimensional spherical scanner 4 are perpendicular to the area to be measured, and the scanning speed is set to 200-400mm / s.

[0027] The present invention can be used to digitally detect various production properties of large-sized metal siding skins, and guide and adjust the large-sized metal siding production process in combination with production standards. Compared with the prior art, the present invention can overcome the difficulties of large siding workpiece size, long manual detection time, and large detection errors, and realize digital detection of large-sized metal siding skin properties based on digital detection technology. In detail:

[0028] First, the present invention establishes a fully covered measurement and tracking field of view, and covers all areas of the large-size metal siding through the tracker, and completely obtains the three-dimensional data and two-dimensional data of the large-size metal siding, which solves the problems of the prior art that large-size metal siding is difficult to measure and has large measurement errors. The present invention significantly enhances its adaptive ability to measure sidings of different sizes, and provides a guarantee for realizing digital measurement of sidings of different sizes.

[0029] Second, the present invention adopts a method combining a three-dimensional spherical scanner 4 and a two-dimensional industrial camera 3 as a means of data acquisition, which solves the limitation of existing methods that they can only measure in two or three dimensions, and at the same time provides rich data support for subsequent measurements of more attributes, and provides a method for the coordination of data of different dimensions and modalities.

[0030] Third, the present invention adopts a series of point cloud and surface quality automated scanning, photographing path planning methods and process parameters of large-size metal wall panel skins, so that the present invention can accurately calculate properties such as wall thickness and rib width, as well as accurately detect surface defects of the skin. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is an overall schematic diagram of the digital detection system for the skin properties of large-size metal sidings of the present invention;

[0032] Figure 2 is a schematic diagram of the full coverage tracking field of view of the present invention;

[0033] Figure 3 It is a schematic diagram of the fixed tooling of the present invention;

[0034] Figure 4 The present invention is a flow chart of the method for digitally detecting the properties of large-size metal wall panels. DETAILED DESCRIPTION

[0035] The specific embodiments of the present invention are further described in detail below with reference to the accompanying drawings.

[0036] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0037] Although the steps in the present invention are arranged with numbers, they are not used to limit the order of the steps. Unless the order of the steps is clearly stated or the execution of a certain step requires other steps as a basis, the relative order of the steps can be adjusted. It is understood that the term "and / or" used in this article involves and covers any and all possible combinations of one or more of the associated listed items.

[0038] The present invention discloses a digital detection scheme for the skin properties of large-sized metal sidings, which aims to digitally detect the skin production parameters of large-sized metal sidings, such as size, wall thickness, rib width, etc., as well as defects on the skin surface through digital three-dimensional detection, two-dimensional detection and other technologies, so as to solve the problems of low measurement efficiency of large-sized workpieces and human operation errors in industrial production, improve the accuracy and efficiency of the measurement process, and further enhance the degree of digitalization of production.

[0039] The idea of ​​realizing the purpose of the present invention is to build a digital detection system for large-sized metal siding skins by using two-dimensional and three-dimensional hardware equipment.

[0040] Figure 1 This is an overall schematic diagram of the digital detection system for large-size metal wall panel skin properties of the present invention. Figure 1The present invention proposes a large-size metal wall panel skin property digital detection system, which includes a binocular optical tracker 1, a fixed tool 2, a two-dimensional industrial camera 3, a three-dimensional spherical scanner 4, a multi-axis robotic arm 5, an automatic guided vehicle 6, a software analysis workstation 7 and a safety net 8.

[0041] The binocular optical tracker 1 achieves full coverage of the measurement and tracking field of view of the large-sized metal siding, and is used to determine the coordinates of the three-dimensional spherical scanner 4. The binocular optical tracker 1 can be multiple groups. The number and position of the binocular optical tracker 1 are determined by the required field of view range of the largest-sized metal siding and other related factors. In the preferred embodiment of the present invention, it can be set to three groups. Figure 1 As shown, the three groups are respectively arranged at the center position of the inner side of the arc of the wall plate to be inspected and at the two ends of the outer side of the arc of the wall plate to be inspected. Figure 2 Schematic diagram of the full coverage tracking field of view of the present invention. Figure 2 As shown, the present invention realizes full coverage tracking field of view through three groups of binocular optical trackers 1.

[0042] The fixed tool 2 is placed at a position set within the field of view determined by the binocular optical tracker 1 and is used to place the wall panel to be inspected. Figure 3 Schematic diagram of the fixing tool 2 of the present invention. In one embodiment, Figure 3 As shown, the fixing fixture 2 is in an arc shape matching the wall panel to be inspected, and has a column at each end for clamping the wall panel to be inspected.

[0043] The two-dimensional industrial camera 3 and the three-dimensional spherical scanner 4 are fixed on the end axis of the multi-axis robot 5 to measure the wall plate to be inspected. The multi-axis robot 5 has multiple axes and can adjust the end position in three dimensions. The automatic guided vehicle 6 and the multi-axis robot 5 are assembled into a composite robot as a measurement platform. The software analysis workstation 7 receives the collected two-dimensional data and three-dimensional data and completes the analysis work. The safety net 8 is used to enclose the inspection station to achieve closed space management.

[0044] In the above-mentioned large-scale metal siding skin property digital detection system of the present invention, the core device includes a binocular optical tracker 1, a composite robot, a three-dimensional spherical scanner 4 and a two-dimensional industrial camera 3. The present invention uses multiple sets of binocular optical trackers 1 to build a set of full-coverage tracking fields of view, and uses a composite robot equipped with a three-dimensional spherical scanner 4 to move and scan in the field of view according to the planned route to obtain three-dimensional point cloud data of the large-scale metal siding skin. At the same time, it is equipped with a two-dimensional industrial camera 3 to obtain two-dimensional image data of the surface of the large-scale metal siding skin. Through the combination of two-dimensional equipment and three-dimensional equipment, a rich data source can be provided to provide support for the realization of diversified detection attributes. The present invention can also transmit the three-dimensional point cloud data and the two-dimensional image data back to the software analysis workstation 7, and use the three-dimensional algorithm and the two-dimensional deep learning algorithm to process the three-dimensional point cloud data and the two-dimensional image data respectively, analyze the properties to be inspected such as the skin wall thickness, rib width and other properties, as well as the defect detection results of the skin surface, and issue a result report.

[0045] The present invention also provides a method for digitally detecting the properties of large-size metal siding skins, which uses the above-mentioned system for digitally detecting the properties of large-size metal siding skins, and includes the following steps: S1, building a fully covered measurement and tracking field of view based on a binocular optical tracker 1; S2, assembling a composite robot; S3, starting the composite robot to follow a planned route, using a carried three-dimensional spherical scanner 4 to obtain three-dimensional point cloud data on the siding surface, and a two-dimensional industrial camera 3 to obtain two-dimensional image data on the siding surface; S4, transmitting data to a software analysis workstation 7 to complete skin property analysis and defect detection; S5, generating a report, and lifting the siding.

[0046] In the digital detection method of the skin property of large-sized metal sidings of the present invention, first, a full-coverage measurement and tracking field of view based on a binocular optical tracker 1 is built to achieve full field coverage of large-sized metal sidings; secondly, an automatic guided vehicle 6 and a multi-axis mechanical arm 5 are assembled into a composite robot, and a three-dimensional spherical scanner 4 and a two-dimensional industrial camera 3 are installed on the multi-axis mechanical arm 5; then, when the siding to be tested is detected, the composite robot is started, and according to the planned route, the three-dimensional spherical scanner 4 is used to obtain three-dimensional point cloud data, and the two-dimensional industrial camera 3 is used to obtain two-dimensional data of the siding surface image; finally, the obtained data is transmitted to the software analysis workstation 7, and after being analyzed by the software algorithm, a report of the detection result attributes is output. Thus, the present invention realizes the digitization of the detection process, reduces manual intervention, and improves the accuracy of the measurement process and production efficiency.

[0047] Figure 4 The flowchart of the digital detection method of the large-size metal wall panel skin property of the present invention is shown in FIG. Figure 4 , the implementation steps of the embodiment of the present invention are further described in detail. The digital detection method of the large-size metal siding skin property of the present invention comprises the following steps:

[0048] Step S1, building a full coverage measurement and tracking field of view based on the binocular optical tracker 1;

[0049] In one embodiment, step S1 comprises the following steps:

[0050] Step S101, perform scanning field of view analysis and calculation according to the maximum size of the wall panel to be measured, equipment attributes and site environment, wherein the site environment is required to be a general industrial indoor environment, and direct sunlight should be avoided as much as possible. The ground is required to be relatively flat and the working path should be free of grooves. According to the maximum coverage range of the binocular optical tracker 1, the motion range of the automatic guided vehicle 6 and the multi-axis robotic arm 5, the effective distance of the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3, the system footprint is reasonably set, and the field of view coverage is calculated in detail to achieve full coverage of the measurement and tracking field of view of large-size metal wall panels.

[0051] Step S102: according to the analysis results, combined with the system footprint, field of view and other information, build the fixed tooling 2, set the position of the binocular optical tracker 1, complete the construction of the tracking field of view, and realize the measurement and tracking field of view required for digital detection.

[0052] Step S2, assembling the automatic guided vehicle 6, the multi-axis robotic arm 5, etc. into a composite robot, and installing the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3.

[0053] In one embodiment, step S2 comprises the following steps:

[0054] Step S201, select the multi-axis robot arm 5 and the automatic guided vehicle 6 for assembly, and build a measurement bearing platform, wherein the weight of the equipment to be carried by the automatic guided vehicle 6 and the height of the multi-axis robot arm 5 need to be reasonably analyzed to complete the construction of the bearing platform.

[0055] Step S202: Use a support fixture on the multi-axis robotic arm 5, and install a three-dimensional spherical scanner 4 and a two-dimensional industrial camera 3 on the two arms respectively.

[0056] Step S3: Start the composite robot and obtain three-dimensional data and two-dimensional data according to the planned route.

[0057] In one embodiment, step S3 comprises the following steps:

[0058] Step S301, plan the corresponding route according to the wall panel model, including the movement of the multi-axis robot arm 5, wherein the automatic guided vehicle 6 adopts laser navigation, and manually and reasonably plans the driving route and the movement of the multi-axis robot arm 5 according to the wall panel model, including the relevant three-dimensional scanning action route and two-dimensional shooting action route.

[0059] Among them, when planning the driving path, it is necessary to take into account the obstacles that may exist around the multi-axis robot 5, the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3. By setting a suitable safety distance or performing collision detection, ensure that the path planning can avoid obstacles and ensure safe operation. The optimal scanning distance between the three-dimensional spherical scanner 4 and the wall panel is 300±50mm. Based on the path planning results, the path points for the movement of the multi-axis robot 5 are generated. Each path point contains the position and posture information of the multi-axis robot 5, as well as the position and orientation information of the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3.

[0060] Among them, when performing the three-dimensional scanning action route and the two-dimensional shooting action route, the starting point and the end point of the multi-axis robot 5, the three-dimensional spherical scanner 4 and the two-dimensional industrial camera 3 are first determined to ensure that the entire wall panel surface is covered. The starting point can be selected at one end of the wall panel, and the end point can be selected at the other end of the wall panel. The wall panel surface is divided into multiple small areas for more detailed scanning. The wall panel can be divided according to the geometric shape to ensure that each small area can be scanned by the three-dimensional spherical scanner 4 to ensure the integrity of the product point cloud data.

[0061] For each small area, a scanning path is defined to cover all points in the area. The following strategies can be used to define the scanning path:

[0062] When the surface of the wall panel to be tested is a smooth surface, a serpentine scanning path is used for the area within the smooth surface: starting from the starting point, scanning along the horizontal direction, and then when reaching the boundary, moving horizontally to the next line and changing the scanning direction, and so on, until reaching the end point. At the same time, during scanning, the trajectory operation parameters of the three-dimensional spherical scanner 4 should be ensured to be: the distance from the area to be tested is within the optimal scanning distance (300±30mm), the binoculars of the three-dimensional spherical scanner 4 should be as perpendicular to the area to be tested as possible, and the scanning speed is set to 200-400mm / s.

[0063] When the surface to be measured of the wall panel is a mesh surface with ribs, the rib mesh scanning path is used for the ribs: the rib mesh veins in the small area are used as the scanning path, and the viewing angle of the three-dimensional spherical scanner 4 is adjusted to ensure that the side part of the rib is also completely scanned, and the binoculars of the three-dimensional spherical scanner 4 should be at an angle of 45°±10 with the ribs to be scanned. Scan the left and right sides of the rib mesh in the order of rows and columns, and then move to the next row or column in the vertical direction, and so on, until all the rib meshes are scanned. The bottom surface of the mesh can adopt the same scanning path as the smooth surface.

[0064] Step S302, start the composite robot and execute it twice according to the planned route to obtain three-dimensional point cloud data and two-dimensional image data respectively. When scanning, the scanning parameters of the three-dimensional spherical scanner 4 are set to resolution 0.8 and exposure 3.0. When taking pictures, the shooting parameters of the two-dimensional industrial camera 3 are set to exposure 80.

[0065] Step S4: The data is transmitted to the software analysis workstation 7 for algorithm analysis.

[0066] In one embodiment, step S4 comprises the following steps:

[0067] Step S401, the 3D point cloud data and the 2D image data are transmitted to the workstation via a wireless network, wherein network parameters such as network bandwidth and frame size need to be considered to optimize the transmission speed;

[0068] Step S402, using a three-dimensional point cloud algorithm to calculate wall thickness, rib width and other properties of the wall panel;

[0069] Among them, the processes and algorithms involved are: first, data preprocessing, using the preprocessing filtering algorithm to remove outliers and noise points from the point cloud data; then surface reconstruction, using the surface reconstruction algorithm to convert the point cloud data into a three-dimensional model or mesh data of the wall panel; finally, the calculation is completed, and the properties such as the wall panel thickness and rib width are calculated according to the relevant features.

[0070] Data preprocessing: The acquired point cloud data needs to be preprocessed to remove noise and invalid points. The relevant preprocessing steps include outlier filtering, smoothing filtering, etc. Outlier filtering can identify and remove outliers by calculating the distance of each point's neighborhood points. Smoothing filtering can use methods such as Gaussian filtering or moving average filtering to reduce noise. Sampling can reduce the density of point cloud data to reduce the amount of calculation.

[0071] The specific algorithm parameters of data preprocessing need to be set reasonably to maintain the optimal detection effect.

[0072] The formula for smoothing filtering is as follows:

[0073]

[0074] Among them, f(x,y,z) represents the value of a point in the original point cloud data, f'(x,y,z) represents the value of the filtered point, σ represents the standard deviation of the Gaussian filter, (i,j,k) is a variable used to represent the neighborhood range of the Gaussian filter, representing the coordinate offset in three-dimensional space;

[0075] If the 3D point cloud data of the wall panel comes from multiple scanning positions or angles, point cloud registration is required to align them to the same coordinate system. Point cloud registration can be achieved using the ICP (Iterative Closest Point) algorithm or other registration algorithms. During the registration process, the position and posture of the point clouds are optimized by calculating the minimum distance and minimum square error between the point clouds to align them.

[0076] Before surface reconstruction, a nearest neighbor search is required to obtain the neighborhood information of each point. The nearest neighbor search can be implemented using data structures such as kd trees and octrees to improve search efficiency. Through the nearest neighbor search, the neighborhood point set of each point can be obtained for subsequent surface reconstruction and thickness calculation.

[0077] Surface reconstruction: After obtaining the domain information of the points, the surface reconstruction algorithm is used to convert the point cloud data into a three-dimensional model or mesh data of the wall panel.

[0078] Commonly used surface reconstruction algorithms include triangulation-based algorithms, such as Delaunay triangulation and Poisson reconstruction, and voxel-based algorithms, such as the Marching Cubes algorithm. These algorithms generate continuous triangular facets or voxel grids based on the geometric shape and distribution of point cloud data to form a surface model of the wall panel. The specific algorithm type needs to be selected after testing.

[0079] Complete the calculation: Calculate the properties of the panel thickness and rib width based on the 3D model or mesh data of the panel. The common method is to calculate the normal vector of each point on the panel surface, and then calculate the thickness by projecting along the normal vector direction. By calculating the distance of the nearest neighbor of the surface point, the thickness distribution and rib width of the panel can be obtained.

[0080] Step S403: Use a deep learning algorithm to detect scratches, thread damage and other defects on the wallboard surface based on the two-dimensional image data. The present invention uses a deep learning algorithm to construct a neural network model using an attention mechanism of morphological topological relationships. The network can model the morphological topological features of scratches, thread damage and other defects, thereby achieving accurate detection of wallboard surface defects.

[0081] Step S5, output the test result report and load and unload materials.

[0082] In one embodiment, step S5 comprises the following steps:

[0083] Step S501: After algorithm analysis, the results of various properties of the skin are obtained, and the software outputs a test report according to the production standard.

[0084] The algorithm analyzes and calculates the properties of the skin, such as thickness and rib width, as well as the defects on the skin surface. The software outputs a skin property measurement report and a skin surface defect location map to guide subsequent production optimization.

[0085] Step S502: remove the current wallboard and import a new wallboard to be inspected.

[0086] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

[0087] It should be noted that the above is only an illustrative description and explanation of the present invention, and those skilled in the art should understand that any modification and replacement of the present invention belongs to the protection scope of the present invention.

Claims

1. A large-size metal siding skin property digital detection system, characterized in that: The invention comprises a binocular optical tracker (1), a fixed tool (2), a two-dimensional industrial camera (3), a three-dimensional spherical scanner (4), a multi-axis mechanical arm (5), an automatic guided vehicle (6), a software analysis workstation (7) and a safety net (8). The binocular optical tracker (1) realizes full coverage of the measurement and tracking field of view of large-sized metal wall panels and is used to determine the coordinates of the three-dimensional spherical scanner (4). The fixed tool (2) is placed at a position set within the field of view determined by the binocular optical tracker (1) and is used to place the wall panels to be inspected. The two-dimensional industrial camera (3) and the three-dimensional spherical scanner (4) are fixed on the end axis of the multi-axis mechanical arm (5) and are used to measure the wall panels to be inspected. The automatic guided vehicle (6) and the multi-axis mechanical arm (5) are assembled into a composite robot as a measurement bearing platform. The software analysis workstation (7) receives the collected two-dimensional data and three-dimensional data and completes the analysis work. The safety net (8) is used to enclose the inspection station and realize the closed space management.

2. The large-size metal siding skin property digital detection system according to claim 1 is characterized in that: The binocular optical trackers (1) are arranged in three groups, and the three groups are respectively arranged at the center position of the inner side of the arc of the wall plate to be inspected and at the two ends of the outer side of the arc of the wall plate to be inspected.

3. The large-size metal siding skin property digital detection system according to claim 1 is characterized in that: The fixing tool (2) is in an arc shape matching the wall plate to be inspected, and has a column at each end for clamping the wall plate to be inspected.

4. The large-size metal siding skin property digital detection system according to claim 1 is characterized in that: The multi-axis robot arm (5) has multiple axes and can adjust the end position in three-dimensional directions.

5. A method for digitally detecting the properties of large-size metal siding skins, using the digital detection system for large-size metal siding skins according to claim 1, characterized in that: The method comprises the following steps: S1, constructing a full-coverage measurement and tracking field of view based on a binocular optical tracker (1), installing the wall panel to be measured on a fixed tooling (2), and realizing full-view coverage of the wall panel to be measured; S2, assembling an automatic guided vehicle (6) and a multi-axis mechanical arm (5) into a composite robot, and installing a three-dimensional spherical scanner (4) and a two-dimensional industrial camera (3) on the multi-axis mechanical arm (5); S3, starting the composite robot, and according to the planned route, using the three-dimensional spherical scanner (4) to obtain three-dimensional point cloud data of the wall panel surface, and using the two-dimensional industrial camera (3) to obtain two-dimensional image data of the wall panel surface; S4, transmitting the obtained data to a software analysis workstation (7), completing skin property analysis and defect detection; S5, generating a detection result property report, and evacuating the wall panel to be measured.

6. The digital detection method for large-size metal siding skin properties according to claim 5 is characterized in that: Step S1 comprises the following steps: Step S101: perform scanning field of view analysis and calculation according to the maximum size of the wall panel to be measured, the equipment attributes and the on-site environment; reasonably set the system footprint according to the maximum coverage of the binocular optical tracker (1), the motion range of the automatic guided vehicle (6) and the multi-axis mechanical arm (5), the effective distance of the three-dimensional spherical scanner (4) and the two-dimensional industrial camera (3); perform detailed calculation of the field of view coverage, and achieve full coverage of the measurement and tracking field of view of the wall panel to be measured; Step S102: according to the analysis results, the fixed tooling (2) is constructed, and the position of the binocular optical tracker (1) is set to complete the construction of the tracking field of view, thereby realizing the measurement and tracking field of view required for digital detection.

7. The digital detection method for large-size metal siding skin properties according to claim 5 is characterized in that: Step S2 comprises the following steps: Step S201: assemble the multi-axis robot arm (5) and the automatic guided vehicle (6) to build a measurement bearing platform, wherein the weight of the equipment to be carried by the automatic guided vehicle (6) and the height of the multi-axis robot arm (5) are analyzed to complete the construction of the bearing platform; Step S202: Use a support fixture on the multi-axis robotic arm (5), and install a three-dimensional spherical scanner (4) and a two-dimensional industrial camera (3) on the two arms respectively.

8. The digital detection method for large-size metal siding skin properties according to claim 5 is characterized in that: Step S3 includes the following steps: Step S301: According to the model of the wall panel to be tested, the automatic guided vehicle (6) adopts laser navigation, manually and reasonably plans the driving route and the movement of the multi-axis robot (5), and generates path points for the movement of the multi-axis robot (5) according to the path planning results, each of which contains the position and posture information of the multi-axis robot (5), as well as the position and orientation information of the three-dimensional spherical scanner (4) and the two-dimensional industrial camera (3); Step S302: Start the composite robot and execute it twice according to the three-dimensional scanning action route and the two-dimensional shooting action route respectively to obtain three-dimensional point cloud data and two-dimensional image data.

9. The digital detection method for large-size metal siding skin properties according to claim 5 is characterized in that: Step S4 comprises the following steps: Step S401, the three-dimensional point cloud data and the two-dimensional image data are transmitted to the workstation via a wireless network; Step S402, using a three-dimensional point cloud algorithm to calculate the properties of the wall thickness and rib width based on the three-dimensional point cloud data; Step S403: Use a deep learning algorithm to detect scratches and thread defects on the wall panel surface based on the two-dimensional image data.

10. The digital detection method for large-size metal siding skin properties according to claim 8 is characterized in that: In step S301: when planning the driving path, possible obstacles around the multi-axis robot (5), the three-dimensional spherical scanner (4) and the two-dimensional industrial camera (3) are considered, and the path planning is ensured to avoid obstacles by setting a suitable safety distance or performing collision detection; firstly, the starting point and the end point of the multi-axis robot (5), the three-dimensional spherical scanner (4) and the two-dimensional industrial camera (3) are determined to ensure that the entire wall panel surface is covered; the wall panel surface is divided into a plurality of small areas for more detailed scanning, and for each small area, a scanning path is defined to cover all points in the area, and the definition of the scanning path adopts the following strategy: When the surface of the wall panel to be tested is a smooth surface, a serpentine scanning path is used for the area inside the smooth surface: starting from the starting point, scanning along the horizontal direction, and then when reaching the boundary, moving to the next line and changing the scanning direction, and so on, until reaching the end point. The trajectory operation parameters of the three-dimensional spherical scanner (4) are as follows: the distance from the area to be measured is within the optimal scanning distance of 300±30mm, the binoculars of the three-dimensional spherical scanner (4) are perpendicular to the area to be measured, and the scanning speed is set at 200-400mm / s.

Images