Three-dimensional scanning system and method for composite material component

Abstract

The invention discloses a three-dimensional scanning system and method for a composite material component. The method comprises the steps that an industrial robot drives a three-dimensional scanner to move to the surface of the composite material component to obtain local three-dimensional point cloud data; a first mark point is arranged on the surface of the three-dimensional scanner, and a second mark point is arranged around the composite material component; the upper computer generates an optimal tracking path in real time according to the first mark point and the second mark point; and guiding the robot to adjust the pose of the laser tracker according to the optimal tracking path so as to maintain the tracking of the three-dimensional scanner by the laser tracker. According to the laser tracker, the three-dimensional scanner and marking points placed around the workpiece can be automatically tracked, the planned motion path is combined with the motion of the guide robot, the guide robot drives the laser tracker to automatically move, the assembly positioning precision of the large composite material stringer on the wall plate skin is detected, and the detection precision of the large composite material stringer on the wall plate skin is improved. And the automatic manufacturing degree of the large composite material component is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automated manufacturing of large-scale composite materials for aerospace, and specifically to a three-dimensional scanning system and method for composite materials, and in particular to a three-dimensional scanning system and method for detecting the assembly positioning accuracy of composite stiffened wall panels. Background Art

[0002] With the continuous improvement of manufacturing standards, higher quality requirements are being placed on large-scale workpieces manufactured in fields such as aerospace, automotive, and shipbuilding. In the aerospace field, the inspection of the positioning accuracy of the long stringer assembly during the manufacturing process of large composite wall panels is extremely critical. However, due to the large size of these workpieces, traditional contact measurement methods, such as three-dimensional coordinate measuring machines, are unable to complete the entire measurement of large-scale workpieces and have low measurement efficiency.

[0003] With the continued maturity of non-contact measurement methods, represented by 3D vision measurement technology, structured light scanning-based 3D measurement methods have emerged to address the problem of workpiece dimensional measurement. By scanning the workpiece's surface to capture 3D point cloud data of the workpiece's shape, the point cloud data is then measured using 3D measurement software. However, due to the limited measurement range of structured light scanners, it is impossible to obtain complete point cloud data for large workpieces that exceed the scanner's field of view in a single scan. Therefore, tracking-based 3D scanning technology, based on dynamic optical tracking, has emerged to address the point cloud acquisition problem for large workpieces, while meeting the requirements of industrial sites in terms of both accuracy and efficiency. Tracking-based structured light 3D scanning technology utilizes a structured light scanner as its core component, along with a laser tracker and 3D point cloud data processing and measurement software. The laser tracker tracks markers on the structured light scanner, registers the point cloud data acquired at different positions, and stitches together a complete 3D point cloud of the large workpiece. By expanding the working range of the structured light scanner, this solves the problem of non-contact dimensional measurement of large workpieces, ultimately enabling quality monitoring during the machining process of large workpieces.

[0004] In large-scale 3D spatial measurement systems based on structured light scanning, achieving tracking of the structured light scanner by a laser tracker is a key issue affecting measurement accuracy and efficiency. Typically, this involves placing the laser tracker on a moving device such as a mobile robot, and designing a sound motion control method. This method, along with a tracking control method, enables the mobile robot to move and dynamically track the structured light scanner. Ultimately, while ensuring the registration accuracy of the large-scale 3D spatial point cloud data acquired by the structured light scanner and the efficiency of automated measurement, a complete large-scale 3D spatial measurement system is constructed, enabling non-contact 3D measurement of large workpieces.

[0005] When using a structured light scanner to collect 3D point cloud data for large workpieces, the limited field of view of the structured light scanner prevents it from covering the entire workpiece. Therefore, a laser tracker is required to track the structured light scanner and dynamic markers, achieving registration of the workpiece point cloud data acquired in different regions, thereby obtaining complete 3D point cloud data for the workpiece. Patent 201810565392.4 proposes a scanning method and system for a tracking-based 3D scanning device, and outlines the basic principles and methods of tracking-based 3D scanning. To enable the laser tracker to track the 3D scanner, it is necessary to ensure that the laser tracker observes at least three markers to complete the pose estimation of the structured light scanner. Patent 202310402014.5 proposes a 3D scanning method and system based on a tracking scanning system, and provides the conditions under which the laser tracker can track the structured light scanner, thereby ensuring that the markers on the structured light scanner are always captured by the laser tracker during the scanning process. Alternatively, a large workpiece can be simply divided into sections according to the structured light scanner's field of view. After scanning each area of the workpiece, manual observation is used to confirm that the structured light scanner is within the laser tracker's field of view. The laser tracker is then moved to the next location manually or through a motion mechanism, and the structured light scanner is used to collect point cloud data from the workpiece. This processing method does not provide a specific motion control method for the laser tracker. It only determines the position of the laser tracker in advance through manual observation, and then tracks the structured light scanner and the markers.

[0006] Existing tracking methods firstly affect the degree of automation in point cloud data acquisition for large-scale workpieces. Before each 3D scan of a new workpiece, the laser tracker's motion needs to be readjusted. More importantly, manually determining whether the laser tracker's tracking range covers the structured light scanner's scanning range is overly conservative. Theoretically, as long as there are enough markers on the structured light scanner that fall within the laser tracker's tracking range, and the laser tracker can acquire enough dynamic markers arranged around the workpiece, it can track the structured light scanner. Finally, tracking by setting a simple motion path for the laser tracker cannot guarantee the tracking quality of the markers within the laser tracker's field of view. Once a tracking error occurs in a marker, the accuracy of the entire workpiece's 3D point cloud data registration is reduced. Alternatively, due to an insufficient number of tracked markers or positional singularities in the tracked markers, point cloud data registration may fail. Summary of the Invention

[0007] Based on this, in order to solve the technical problems existing in the prior art, the present invention provides a three-dimensional scanning system and method for composite material components, in particular, a three-dimensional scanning system and method for detecting the assembly positioning accuracy of composite material reinforced wall panels.

[0008] The present invention provides a three-dimensional scanning system for composite material components, comprising an industrial robot, a three-dimensional scanner, a laser tracker, a guide robot and a host computer; The industrial robot is connected to the three-dimensional scanner, and the industrial robot is configured to drive the three-dimensional scanner to move to multiple local scanning positions on the surface of the composite material component to scan and obtain local three-dimensional point cloud data of the composite material component; The three-dimensional scanner surface is provided with a plurality of first marking points, and the composite material component is surrounded by spatially distributed second marking points; the host computer is configured to select a first marking point and a second marking point that meet an optimization goal from the plurality of first marking points and the plurality of second marking points, and to generate an optimal tracking path in real time based on the first marking point and the second marking point that meet the optimization goal; the optimization goal is to select the plurality of first marking points and the second marking point so that the position singularity of the laser tracker is minimized; The laser tracker is connected to the guide robot to form a linkage motion mechanism. The guide robot is configured to adjust the position of the laser tracker according to the optimal tracking path to keep the laser tracker tracking the first marking point on the surface of the three-dimensional scanner.

[0009] Furthermore, the host computer is also configured to establish a dynamic mapping relationship between the local three-dimensional point cloud data of the composite material component and its spatial distribution based on the spatial coordinates of the first marking point and the second marking point, and splice the local three-dimensional point cloud data according to the dynamic mapping relationship to obtain an overall three-dimensional model of the structure.

[0010] The present invention provides a three-dimensional scanning method for a composite material component, comprising: Obtaining first marking points on the surface of the three-dimensional scanner for being tracked by a laser tracker, and second marking points set around the scanned composite material component for providing positioning reference information to the laser tracker; generating an optimal tracking path for guiding the robot based on the first marker point and the second marker point; wherein the first marker point and the second marker point are the first marker point and the second marker point that satisfy an optimization goal, wherein the optimization goal is to select a plurality of first marker points and the second marker point so as to minimize the position singularity of the laser tracker; Based on the optimal tracking path, the robot is driven to adjust the position of the laser tracker to keep the laser tracker tracking the 3D scanner.

[0011] Furthermore, the first marker point and the second marker point are the first marker point and the second marker point that meet an optimization goal, and the optimization goal is: selecting a number of first marker points and second marker points so that the position singularity of the laser tracker is minimized, specifically including: Build a marker optimization model: Constructing the homogeneous transformation matrix of the global coordinate system : in, Represents the rotation matrix elements of the global coordinate system, which together constitute the first rotation matrix , Respectively represent the translation of the three coordinate axes of the global coordinate system, together forming the first translation vector ; Get the first i The homogeneous transformation matrix of the marker points : Where, Indicates the i The rotation matrix elements of the marker points together constitute the second rotation matrix , Respectively represent i The translation of the three coordinate axes of the marker points together constitute the second translation vector ; get and The Euclidean distance between , angle difference and relative error : in, and The quaternions representing the two poses respectively; is the identity matrix; According to the Euclidean distance , angle difference and relative error , build a marker optimization model: Where, is an indicator variable, with a value of 0 or 1. is the total number of data points in the point cloud; The optimization model is solved using a branch and bound method to obtain a plurality of first marking points and a plurality of second marking points that meet the optimization objective.

[0012] Furthermore, generating an optimal tracking path for guiding the robot based on the first marking point and the second marking point specifically includes: Construct a motion coordinate system and obtain a motion model for describing the changes of the guide robot, the first marker, and the second marker in the motion coordinate system: in, and Respectively represent the guiding robot in the Moment and The state vector at the moment; , is the position vector of the guiding robot, is a marker feature vector consisting of the coordinates of the first marker and the second marker; is the equation of state; is the control quantity that guides the robot; is a zero-mean, constant covariance matrix Gaussian random variable; is Gaussian noise with zero mean; Is to guide the robot The observation vector at time instant; represents the observation equation; represents Gaussian noise with zero mean; The nonlinear motion model is linearized by using the extended Kalman filter, and the standard Kalman filter is applied to the linearized motion model for state estimation to generate the optimal tracking path.

[0013] Furthermore, the driving and guiding robot to adjust the position and posture of the laser tracker based on the optimal tracking path specifically includes: The motion model is designed based on the Lyapunov stability method to ensure that the error between the actual position and actual direction of the guiding robot and the preset position and preset direction in the optimal tracking path is Minimum; the motion model is: in, It is a guide robot Actual speed in direction; It is a guide robot Actual speed in direction; is the actual angular velocity of the guided robot; is the desired linear velocity of the guiding robot; is the linear velocity of the guided robot; is the actual direction angle of the guiding robot; It is a guide robot The actual position of the direction; It is a guide robot The actual position of the direction; is the desired angular velocity to guide the robot; is the angular velocity of the guided robot.

[0014] At least one of the above technical solutions adopted by the present invention can achieve the following beneficial effects: In the three-dimensional scanning system and method for composite components provided by the present invention, by designing a laser tracker motion control method, the laser tracker can realize autonomous tracking of the three-dimensional scanner and the marker points placed around the workpiece, ensuring that the point cloud data of the composite component collected in different areas can be accurately aligned, providing a guarantee for the construction of the three-dimensional space measurement system; the planned motion path is combined with the motion of the guide robot to realize the autonomous motion of the laser tracker driven by the guide robot, thereby improving the degree of automation of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0016] Figure 1 A partial schematic diagram of the three-dimensional scanning system for composite material components provided by the present invention; Figure 2 A flow chart of the laser tracker motion control method provided by the present invention; Figure 3 Flowchart of the branch and bound algorithm for solving the marking point optimization problem provided by the present invention; Figure 4 The appearance of the workpiece to be measured provided by the present invention; Figure 5 Point cloud data collected by the three-dimensional space measurement system provided by the present invention; Figure 6 This is an overall schematic diagram of the three-dimensional scanning system for composite material components provided by the present invention. DETAILED DESCRIPTION

[0017] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with specific embodiments of the present invention and corresponding drawings. Obviously, the embodiments described are only some embodiments of the present invention, not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of the present invention.

[0018] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0019] The three-dimensional scanning method of a composite material component provided in the embodiment of the present application can be applied to Figure 1 and Figure 6 The composite component 3D scanning system shown in FIG. 1 includes an industrial robot, a 3D scanner, a laser tracker, a guide robot, and a host computer. The industrial robot is connected to the 3D scanner and is configured to drive the 3D scanner to move to the composite component (i.e. Figure 2Multiple local scanning positions on the surface of a workpiece (in the 3D scanner) are used to scan and acquire local 3D point cloud data of the composite component. A plurality of first markers are provided on the surface of the 3D scanner, and second markers are spatially distributed around the composite component. A host computer is configured to generate an optimal tracking path in real time based on the first and second markers. The host computer is further configured to establish a dynamic mapping relationship between the local 3D point cloud data of the composite component and its spatial distribution based on the spatial coordinates of the first and second markers, and to splice the local 3D point cloud data based on the dynamic mapping relationship to obtain an overall 3D model of the structure. A laser tracker is connected to a guide robot to form a linked motion mechanism. The guide robot is configured to adjust the position of the laser tracker according to the optimal tracking path to maintain tracking of the first marker on the surface of the 3D scanner. The core component is a 3D scanner, which is used to acquire 3D point cloud data of the workpiece. This device is mounted on the end of an industrial robot and uses the robot's movements to scan point cloud data at different locations on the workpiece. The laser tracker, a tracking device such as a 3D scanner, is used to observe multiple markers distributed on the 3D scanner and around the workpiece to track the position of the 3D scanner. The position and posture of multiple marker points are used to realize the registration of the local point cloud data of the workpiece collected by the 3D scanner, thereby obtaining a complete 3D point cloud of the workpiece. For composite components, when the 3D scanner collects the point cloud data of the workpiece, the size of the workpiece usually exceeds the field of view of the laser tracker, resulting in the laser tracker being unable to obtain the position and posture of the marker points of the 3D scanner outside the field of view. In order to solve the above problems, on the one hand, it is necessary to ensure the tracking of the 3D scanner through the movement of laser tracking, and on the other hand, it is necessary to arrange dynamic marker points around the workpiece and obtain the position and posture of these marker points in advance to ensure that the laser tracker can always observe the 3D scanner and obtain a sufficient number of dynamic marker points, so as to achieve accurate registration of the point cloud data of composite components in different postures.

[0020] Therefore, to achieve 3D point cloud data scanning of composite components, the following key issues need to be addressed: how to use an automated guided vehicle (AGV) to drive a laser tracker to track the markers on the 3D scanner and the dynamic markers placed around the workpiece, thereby ensuring accurate registration of the composite component's point cloud data. This paper decomposes the 3D scanning problem of structural components into three subproblems: marker selection, laser scanner motion path generation, and AGV motion control. First, to improve point cloud registration accuracy and algorithm real-time performance, an optimization method is used to select markers, using the pose singularities and number of the 3D scanner markers and the markers placed around the workpiece within the laser tracker's field of view as performance indicators. Second, to plan the motion path of the AGV driving the laser tracker to track the markers, a laser tracker position estimation method based on an extended Kalman filter is proposed, using the markers obtained in the first problem as beacons. Third, to enable the AGV to track the planned path, a dynamic tracking control method for a differentially driven AGV based on a reference path is proposed, combining the AGV's kinematic model and the nonholonomic constraints it faces.

[0021] To address the aforementioned issues, the laser tracker control method proposed in this invention solves the problem of structured light scanning of composite components. The laser tracker, through the motion of an automated guided vehicle (AGV), tracks the markers on a 3D scanner and those placed around the workpiece. This ensures accurate registration of the workpiece point cloud data acquired by the 3D scanner through continuous scanning, ensuring the accuracy of the collected workpiece point cloud data while also improving the automation level of the entire system. The laser tracker motion control method for 3D spatial measurement consists of three steps, which are executed sequentially until the entire workpiece is scanned. In each step, the problem is first described, followed by a mathematical model, and finally a solution method is provided. First, the markers to be tracked are selected, their mathematical model is established, and the solution is obtained by selecting an appropriate optimization method. The specific modeling approach and solution method are referenced in the literature [Li D, Sun X. Nonlinear integer programming [M]. Springer Science & Business Media, 2006]. Then, the laser tracker position estimation method based on extended Kalman filter and the AGV path planning method are used to plan the path of the AGV carrying the laser tracker. This problem can be equivalent to the SLAM problem in the field of mobile robots. Its modeling and solution refer to the method in the literature [Shi Ye, Wu Huaiyu, Xu Wenxia, et al. Research on mobile robot SLAM based on extended Kalman filter [J]. Electronic Design Engineering, 2012, 20(1):3.]. Finally, the motion control method of differential drive AGV based on feedback linearization is used to design the controller for the AGV. The common method of mobile robot control trajectory tracking is referred to [Xiaoping Y, Yamamoto Y. Dynamic feedback control of vehicles with two steerable wheels [C] / / . Proceedings of the International Conference on Robotics and Automation. Minneapolis, MN, USA: IEEE, 1996, 12(1), 1006-1010], and the AGV drives the laser tracker to track the structured light scanner. Combining the above methods solves the motion control problem of the laser tracker for three-dimensional spatial measurement, and can provide accurate point cloud data registration and efficient automated measurement for the three-dimensional spatial measurement system based on structured light scanning.

[0022] The following specific Figure 2 The three-dimensional scanning method of a composite material component provided by the present invention is described in detail, and specifically comprises the following steps: S1: Select the marker point within the current field of view of the laser tracker.

[0023] To ensure that the laser tracker can estimate its new pose after movement based on the markers on the 3D scanner and the markers placed around the workpiece, it is necessary to select appropriate markers for pose calculation while ensuring computational efficiency and accuracy. This requires selecting the fewest markers with the least pose singularity from among the many markers. This problem can be formulated as a standard integer linear programming problem. The difference between the pose homogeneous matrix representing the markers and the pose matrix of the global coordinate system is used as a performance indicator. This difference is measured using three metrics: the Euclidean distance, the angular difference, and the relative error between the two poses.

[0024] Assume that the homogeneous transformation matrix of the global coordinate system of the measurement system is Expressed as: ; (1) Where, Represents the rotation matrix elements of the global coordinate system, which together constitute the first rotation matrix , Respectively represent the translation of the three coordinate axes of the global coordinate system, together forming the first translation vector .

[0025] No. i The homogeneous transformation matrix of the marker points Expressed as: ; (2) Where, Definition and Similar to, Indicates the i The rotation matrix elements of the marker points together constitute the second rotation matrix , Respectively represent i The translation of the three coordinate axes of the marker points together constitute the second translation vector .

[0026] The Euclidean distance between the two Expressed as: (3)

[0027] The angle difference between the two Expressed as: ; (4) in, , The quaternions representing the two postures are expressed as follows: ; (5) ; (6) ; (7) Where, It is calculated by the elements in the homogeneous transformation matrix (2), 、 and Represents a vector The three components.

[0028] The relative error between the two Expressed as: ; (8) Where, is the identity matrix.

[0029] According to the above three indicators, the optimization model of the marker selection problem is obtained: ; ; (9) Where, is the Euclidean distance difference between the translation vectors of the two homogeneous transformation matrices, is the relative error of the quaternion representation of the pose, is an indicator variable, with a value of 0 or 1. is the total number of data points in the point cloud.

[0030] The performance metric sums the three error metrics for all marker poses to produce a total cost function. The optimization goal is to select the marker with the greatest variance. The first constraint specifies that at least three markers are required for the laser tracker to calculate the 3D scanner's pose. The second constraint determines whether a particular marker should be selected. Together, the performance metric and the constraints form a mathematical model for an integer optimization problem.

[0031] Since the number of markers within the field of view of the laser tracker is usually more than a dozen, the scale of the optimization problem is not large, so the classic branch and bound method is used to solve it. The algorithm flow is as follows: Figure 3As shown in the figure, first determine the relaxed integer programming problem to be solved, then execute step 1: determine whether there is an optimal solution. If so, execute step 2; otherwise, the problem has no solution. Step 2: determine whether the optimal solution is an integer solution. If so, obtain the optimal solution; otherwise, execute step 3. Step 3: Bounding: find the feasible solution LB of the subproblem and determine whether LB is greater than the set threshold v. If so, obtain the optimal solution; otherwise, execute step 4. Step 4: Branching: construct two new linear programming problems and return to step 3 until the optimal solution is obtained. The set of landmark points obtained by the solution serves as the landmark points for path planning in the next step.

[0032] S2: Perform laser tracker path planning based on the selected marker points.

[0033] The AGV is actually a wheeled robot. The process of driving the laser tracker to track the markers can be equivalent to the SLAM problem of a mobile robot. The AGV acts as the mobile robot, and the markers within the field of view of the laser tracker act as beacons in the map. The motion of the AGV and the changes of the markers in the AGV motion coordinate system are described using the following nonlinear random model: ; ; (10) Among them, the state vector Include dimensional AGV position vector and Fixed The feature vector of the dimensional marker ,Right now The dimension of : ; (11) in, dimensional input vector is the control quantity of AGV, dimensional random vector is a zero-mean, constant covariance matrix Gaussian random process. Function represents the sensor model, represents the error and noise. Also assume is a Gaussian process with mean 0.

[0034] Estimate for the current marker Given a set of measurements ,Mode: ; (12) At a given control input Finally, the a priori noise-free estimate of the new position and marker features of the AGV is given. Similarly:

[0035] ; (13) is a priori noise-free estimate of the sensor measurement value.

[0036] In general, and It is nonlinear, and the linearization method, namely the first-order Taylor expansion, is used to obtain: ; ; (14) Where, 、 and is the Jacobian matrix: .

[0037] Assuming that the markers are fixed, their prior estimates are: (15)

[0038] In this way, the dynamic changes of the overall state model of AGV and mark points are: ; (16)

[0039] It can also be written as: ; ; (17) Where: ; .

[0040] After the above linearization of the dynamic changes of the AGV position and the marking points, the linear stability problem can be directly solved using the method for solving it.

[0041] S3: Calculate the AGV motion controller based on the planned path.

[0042] The robot motion control problem is described by a dynamic model, a kinematic model, and nonholonomic constraints, namely: ; ; ; (18) in, is the derivative of the linear velocity of the guiding robot, i.e., acceleration; is the mass of the guiding robot, is the radius of the guide robot's driving wheels, is the left driving wheel torque that guides the robot, is the right driving wheel torque guiding the robot; is the derivative of the robot's angular velocity, i.e., angular acceleration, is the wheelbase of the guide robot, is the wheel spacing of the guide robot. 、 is the control input, 、 、 Represent the derivatives of the position and orientation angle of the guiding robot respectively. The state vector of the guiding robot is expressed as:

[0043] (19)

[0044] The problem to be solved is to keep track of the desired state trajectory: ; (20) And make the error approach zero.

[0045] To this end, the controller is designed using the Lyapunov stability method. To meet the stability requirements, the desired trajectory must satisfy the kinematic equations and nonholonomic constraints, namely:

[0046] . (twenty one)

[0047] In the local (mobile) coordinate system of the wheeled mobile robot The error expressed in , , It is given by: ; (twenty two) and: . (twenty three)

[0048] Differentiating Equations (22) and (23) and considering Equations (18) and (21), we obtain the kinematic model of the error: : The motion model is designed based on the Lyapunov stability method to ensure that the error between the actual position and actual direction of the guiding robot and the preset position and preset direction in the optimal tracking path is Minimum; the motion model is: ; (twenty four) in, It is a guide robot Actual speed in direction; It is a guide robot Actual speed in direction; is the actual angular velocity of the guided robot; is the desired linear velocity of the guiding robot; is the linear velocity of the guided robot; is the actual direction angle of the guiding robot; It is a guide robot The actual position of the direction; It is a guide robot The actual position of the direction; is the desired angular velocity to guide the robot; is the angular velocity of the guided robot; linear velocity and angular velocity is the motion control variable. Obviously, Equation (24) satisfies the kinematics and nonholonomic constraint equations of WMR.

[0049] Therefore, the motion feedback controller will be constructed based on Equation (24) and designed using the Lyapunov stability method. Since the controller here is nonlinear, its structure cannot be determined in advance. Its structure will be determined by the selected Lyapunov function. Here, the following candidate Lyapunov functions are selected:

[0050] (25)

[0051] Differentiating Equation (8) with respect to time yields: (26)

[0052] In order to make , select control input and , such that: (27)

[0053] It is thus determined that: ； (28) (29)

[0054] Obviously, for and , we can get , only when and Therefore, the controller (Eq. (28) and Eq. (29)) can guarantee global asymptotic tracking of the desired trajectory.

[0055] S4: Determine whether the workpiece scanning is completed. If so, end the tracking; otherwise, return to step S1.

[0056] The motion control algorithm of the AGV is implemented by the upper computer, and then sent to the lower computer AGV controller to be converted into an electrical signal that drives the motor movement of the AGV, ultimately realizing the motion control of the AGV.

[0057] The three-dimensional scanning method of a structural part provided by the present invention has the following effects: 1) By designing a laser tracker motion control method, the laser tracker can autonomously track the 3D scanner and the markers placed around the workpiece, ensuring that the point cloud data of the composite component collected in different areas can be accurately aligned, providing a guarantee for the construction of a 3D spatial measurement system.

[0058] 2) The marker point optimization algorithm selects the marker points with the greatest impact on the point cloud data registration accuracy and the least number from a large number of marker points. While ensuring the registration of point cloud data, it improves the computational efficiency of the point cloud registration algorithm, thereby improving the real-time performance of the system.

[0059] 3) By combining the planned motion path with the AGV's kinematic model and nonholonomic constraints, an AGV motion control algorithm is proposed to achieve autonomous motion of the laser tracker driven by the AGV, thereby improving the automation level of the system.

[0060] When applying the three-dimensional scanning method of the structural parts provided by the present invention, in the marker point selection optimization problem, by adjusting the weights of the number of marker points and the singularity of the posture in the performance index, or introducing new performance indicators such as the physical distance between the marker point and the laser tracker, the algorithm can be balanced in terms of real-time performance and solution accuracy, thereby ensuring that the algorithm can be applied to the actual needs of different scenarios. In the laser tracker path planning method, the extended Kalman filter used in this application can be replaced with other position estimation methods according to the characteristics of the scene, thereby improving the applicability of the algorithm in different scenarios. In the AGV motion control problem, the differential drive AGV kinematic model used in this application can be replaced with a unicycle model, a car-like model, a three-wheeled omnidirectional mobile model or a four-wheeled omnidirectional mobile model for motion controller design, thereby improving the algorithm's compatibility with different types of mobile robots.

[0061] The laser tracker motion control method proposed in the invention has been applied in a composite component 3D spatial measurement system. The composite component 3D spatial measurement system consists of a 3D scanner, a laser tracker, an industrial robot, an AGV, and dynamic markers. The system can realize 3D point cloud data collection and dimensional measurement of composite components with a length of 12 meters. The appearance of the workpiece is as follows: Figure 4As shown in the figure, using point cloud data processing software and 3D measurement software, the 3D point cloud data collected from each workpiece region is aligned to obtain a complete 3D point cloud data set for the composite component. After importing the 3D measurement software, the dimensional measurement results for the composite component are obtained, and a measurement report is generated to indicate whether the critical dimensions of the workpiece meet tolerance requirements. During the point cloud data acquisition process, a 3D scanner is mounted on the end of an industrial robot, and the robot's movement captures point cloud data from the workpiece surface. Due to the large size of the workpiece, the 3D scanner's acquisition range cannot fully cover the entire workpiece. Therefore, the industrial robot is mounted on a linear guide parallel to the workpiece. The 3D scanner is driven by external axis motion to collect local point cloud data from each region of the composite component, and then moves to the next position to continue data acquisition. To stitch the workpiece point cloud data obtained from the scanned regions into a complete 3D model, the laser tracker must be able to track the markers on the 3D scanner and the markers placed at fixed locations around the workpiece, thereby achieving accurate alignment of the point cloud data obtained at the scanning locations. Because the industrial robot drives the 3D scanner to translate along the workpiece, the AGV drives the laser tracker to track the position of the 3D scanner. The three-dimensional scanning of structural parts is achieved by the control method proposed in this application. The specific workflow is as follows:

[0062] (1) The industrial robot starts from its initial position and drives the 3D scanner to scan a local area of the composite component to collect point cloud data of the workpiece surface. The laser tracker collects the marker points on the 3D scanner within the field of view and the marker points around the workpiece, and executes the marker point selection algorithm to obtain a set of marker points with the largest difference and the smallest number.

[0063] (2) Plan the motion path of the laser tracker based on the selected marker points; (3) Build the AGV motion controller according to the planned path and send the control signal to the AGV to realize motion control.

[0064] (4) The laser tracker updates the marker points within the field of view and re-executes the first three steps until the scanning of the composite component is completed.

[0065] Through the above workflow, the point cloud data of the composite material component collected in different regions are registered by the data processing software in the host computer to obtain a complete model of the composite material component, such as Figure 5 Finally, 3D measurement software is used to perform denoising, filtering, and patching of the workpiece's point cloud data, and a measurement report is generated. This complete data acquisition and processing process demonstrates the effectiveness of the proposed method.

[0066] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing the relevant hardware using a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes in the above-described method embodiments. Any reference to memory, storage, database, or other media used in the various embodiments provided herein may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, or optical storage. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM can take various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0067] The technical features of the above embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of the present invention.

Claims

1. A three-dimensional scanning system for composite components, characterized in that: Including industrial robots, 3D scanners, laser trackers, guide robots and host computers; The industrial robot is connected to the three-dimensional scanner, and the industrial robot is configured to drive the three-dimensional scanner to move to multiple local scanning positions on the surface of the composite material component to scan and obtain local three-dimensional point cloud data of the composite material component; The three-dimensional scanner surface is provided with a plurality of first marking points, and the composite material component is surrounded by spatially distributed second marking points; the host computer is configured to select a first marking point and a second marking point that meet an optimization goal from the plurality of first marking points and the plurality of second marking points, and to generate an optimal tracking path in real time based on the first marking point and the second marking point that meet the optimization goal; the optimization goal is to select the plurality of first marking points and the second marking point so that the position singularity of the laser tracker is minimized; The laser tracker is connected to the guide robot to form a linkage motion mechanism. The guide robot is configured to adjust the position of the laser tracker according to the optimal tracking path to keep the laser tracker tracking the first marking point on the surface of the three-dimensional scanner.

2. The three-dimensional scanning system for composite material components according to claim 1, wherein: The host computer is also configured to establish a dynamic mapping relationship between the local three-dimensional point cloud data of the composite material component and its spatial distribution based on the spatial coordinates of the first marking point and the second marking point, and to splice the local three-dimensional point cloud data according to the dynamic mapping relationship to obtain an overall three-dimensional model of the structure.

3. A three-dimensional scanning method for a composite material component applied to the three-dimensional scanning system for a composite material component according to any one of claims 1 to 2, characterized in that: include: Obtaining first marking points on the surface of the three-dimensional scanner for being tracked by a laser tracker, and second marking points set around the scanned composite material component for providing positioning reference information to the laser tracker; generating an optimal tracking path for guiding the robot based on the first marker point and the second marker point; wherein the first marker point and the second marker point are the first marker point and the second marker point that satisfy an optimization goal, wherein the optimization goal is to select a plurality of first marker points and the second marker point so as to minimize the position singularity of the laser tracker; Based on the optimal tracking path, the robot is driven to adjust the position of the laser tracker to keep the laser tracker tracking the 3D scanner.

4. The three-dimensional scanning method of a composite material component according to claim 3, wherein: The first marker point and the second marker point are the first marker point and the second marker point that meet the optimization goal. The optimization goal is to select a number of first marker points and the second marker point so that the position singularity of the laser tracker is minimized, specifically including: Build a marker optimization model: Constructing the homogeneous transformation matrix of the global coordinate system : in, Represents the rotation matrix elements of the global coordinate system, which together constitute the first rotation matrix , Respectively represent the translation of the three coordinate axes of the global coordinate system, together forming the first translation vector ; Get the first i The homogeneous transformation matrix of the marker points : Where, Indicates the i The rotation matrix elements of the marker points together constitute the second rotation matrix , Respectively represent i The translation of the three coordinate axes of the marker points together constitute the second translation vector ; get and The Euclidean distance between , angle difference and relative error : in, and The quaternions representing the two poses respectively; is the identity matrix; According to the Euclidean distance , angle difference and relative error , build a marker optimization model: Where, is an indicator variable, with a value of 0 or 1. is the total number of data points in the point cloud; The optimization model is solved using a branch and bound method to obtain a plurality of first marking points and a plurality of second marking points that meet the optimization objective.

5. The three-dimensional scanning method of a composite material component according to claim 3, wherein: Generating an optimal tracking path for guiding the robot according to the first marking point and the second marking point specifically includes: Construct a motion coordinate system and obtain a motion model for describing the changes of the guide robot, the first marker, and the second marker in the motion coordinate system: in, and Respectively represent the guiding robot in the Moment and The state vector at the moment; , is the position vector of the guiding robot, is a marker feature vector consisting of the coordinates of the first marker and the second marker; is the equation of state; is the control quantity that guides the robot; is a zero-mean, constant covariance matrix Gaussian random variable; is Gaussian noise with zero mean; Is to guide the robot The observation vector at time instant; represents the observation equation; represents Gaussian noise with zero mean; The nonlinear motion model is linearized by using the extended Kalman filter, and the standard Kalman filter is applied to the linearized motion model for state estimation to generate the optimal tracking path.

6. The three-dimensional scanning method of a composite material component according to claim 3, wherein: The driving and guiding robot to adjust the position and posture of the laser tracker based on the optimal tracking path specifically includes: The motion model is designed based on the Lyapunov stability method to ensure that the error between the actual position and actual direction of the guiding robot and the preset position and preset direction in the optimal tracking path is Minimum; the motion model is: in, It is a guide robot Actual speed in direction; It is a guide robot Actual speed in direction; is the actual angular velocity of the guided robot; is the desired linear velocity of the guiding robot; is the linear velocity of the guided robot; is the actual direction angle of the guiding robot; It is a guide robot The actual position of the direction; It is a guide robot The actual position of the direction; is the desired angular velocity to guide the robot; is the angular velocity of the guided robot.

Images