Micron-level precision double-mechanical-arm measuring device with scene perception and method thereof

Abstract

The invention relates to a micron-order precision double-mechanical-arm measuring device with scene perception and a method thereof, the tail end of a power arm is connected with a multi-degree-of-freedom movement device, the multi-degree-of-freedom movement device is connected with a non-contact measuring device of a measuring unit, and the tail end of a driven measuring arm is connected with a contact measuring head of the measuring unit. The rope penetrates through the fixed pulley, one end of the rope is connected with the connecting piece, and the other end of the rope is connected with the counterweight; a plurality of 3D perception cameras are arranged in the measurement scene to cover the measurement environment; the power arm and the multi-degree-of-freedom motion device are active motion mechanisms, the motion of the power arm and the multi-degree-of-freedom motion device drives the measurement unit to move and drives the driven measurement arm connected with the measurement unit to move, a measurement result is obtained through the driven measurement arm, the counterweight is used for stabilizing the driven measurement arm, the plurality of 3D perception cameras are used for measuring environment perception, and the global calibration object is used for calibrating parameters.

Description

Technical Field

[0001] This invention relates to a scene-aware, micron-level precision dual-arm measuring device and method. Background Technology

[0002] A measuring arm (also known as an articulated coordinate measuring machine) is a widely used device in industrial measurement. Its main function is to move the measuring probe or scanning device at its end, through a multi-joint structure, to a target position to achieve three-dimensional spatial measurement. Traditional measuring arms are typically operated by a manual gripper who moves the measuring unit at the end of the arm to the measurement position, followed by data acquisition via contact or non-contact methods. These types of measuring arms rely entirely on manual operation, requiring highly experienced technicians, and their accuracy is significantly affected by human factors. Measurement efficiency is low, and prolonged gripping can cause operator fatigue, further reducing efficiency and making them unsuitable for continuous operation.

[0003] To reduce the burden of manual operation, some technical solutions attempt to combine industrial robots with measuring units to form automated measuring arm systems. While this method achieves automated movement of the measuring arm, insufficient measurement accuracy remains a problem in practical applications. The absolute positioning accuracy of industrial robots is typically at the millimeter or even centimeter level, which is still insufficient compared to high-precision traditional measuring arms, making it difficult to meet the demands of high-precision measurement. Furthermore, robots themselves lack sensing capabilities, requiring operators to specify the position the end-effector measuring unit needs to reach, thus failing to automate the measurement process.

[0004] In summary, both traditional, manually operated measuring arms and automated measuring arms based on robotic technology have significant shortcomings in terms of measurement efficiency, accuracy, and ease of use, making it difficult to simultaneously meet the demands of high-precision and high-efficiency industrial measurement. Therefore, there is a need to develop dual-arm measuring devices with micron-level precision and scene awareness. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a dual robotic arm measurement device and method with scene awareness and micron-level precision, which aims to solve problems such as low measurement efficiency, inability to perform autonomous measurement, and poor measurement quality.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A dual-arm measuring device with micron-level precision and scene awareness is characterized by comprising a power arm, a driven measuring arm, a measuring unit, a multi-degree-of-freedom motion device, multiple 3D sensing cameras, and a global calibration object for parameter calibration. The measuring unit includes an optical non-contact measuring device and a contact measuring head connected to it. The end of the power arm is connected to the multi-degree-of-freedom motion device, which is connected to the non-contact measuring device. The end of the driven measuring arm is connected to the contact measuring head. A connector is mounted on the driven measuring arm, and a rope passes through a fixed pulley, with one end connected to the connector and the other end connected to a counterweight. Multiple 3D sensing cameras are arranged in the measurement scene to cover the measurement environment. The power arm and the multi-degree-of-freedom motion device are active motion mechanisms; their movement drives the measuring unit, which in turn drives the driven measuring arm connected to the measuring unit. Measurement results are obtained through the driven measuring arm. The counterweight is used to stabilize the driven measuring arm. Multiple 3D sensing cameras are used for environmental perception, and the global calibration object is used to calibrate parameters.

[0008] Furthermore, in the aforementioned dual robotic arm measuring device with scene awareness and micron-level precision, the non-contact measuring device is fixed to the end of the multi-degree-of-freedom motion device via screws, and the contact measuring head is integrated with the non-contact measuring device via a clamp.

[0009] Furthermore, in the aforementioned scene-aware, micron-level precision dual robotic arm measuring device, the non-contact measuring device is a structured light 3D scanner or a laser 3D scanner, and the contact measuring head is a contact measuring head of an articulated arm coordinate machine.

[0010] Furthermore, the aforementioned scene-aware, micron-level precision dual-arm measuring device includes a multi-degree-of-freedom motion device comprising a servo motor, a diaphragm coupling, a ball screw, a screw nut, a telescopic rod, a telescopic rod guide seat, a motor mounting base, a frameless torque motor, a drive shaft, a connecting rod, a linear guide rail, and a base. The servo motor and the telescopic rod guide seat are mounted on the base. The servo motor is connected to the ball screw via the diaphragm coupling. The ball screw engages with the screw nut, which is then connected and fixed to the telescopic rod. The telescopic rod passes through the guide hole of the telescopic rod guide seat and connects to the motor mounting base. The telescopic rod slides with the linear guide rail. The frameless torque motor is mounted within the motor mounting base and is connected to the connecting rod via the drive shaft.

[0011] Furthermore, in the aforementioned scene-aware, micron-level precision dual-arm measuring device, the global calibration object consists of three or more calibration units, distributed in a spherical or planar form, connected by linkages to form a whole. Each calibration unit contains three standard spheres and one ArUco code. The three non-collinear standard spheres construct a stable spatial triangular reference, constraining the spatial rotational and translational degrees of freedom of the driven measuring arm.

[0012] Furthermore, the aforementioned dual robotic arm measuring device with scene awareness and micron-level precision also includes a control system connected to the power arm, the driven measuring arm, multiple 3D sensing cameras, and multiple free motion devices. The control system includes a calibration module, a 3D global perception module, a measurement element extraction module, an intelligent tracking module, a 3D data alignment module, a path planning module, a motion and collision avoidance detection module, and a measurement module.

[0013] The measurement element extraction module, based on the large language model, extracts measurement elements of spare parts, extracts information from the input CAD model of the object to be measured, and outputs the features to be measured.

[0014] The calibration module calibrates and calibrates the parameters of the measuring equipment, calibrates the optical-based non-contact measuring device in the measuring unit, calibrates the driven measuring arm, calibrates the coordinate system transformation relationship of the equipment, i.e., the coordinate system transformation matrix, and constructs the global coordinate system.

[0015] The 3D global perception module acquires spatial environment information by deploying multiple 3D perception cameras in the measurement scene. The 3D spatial information acquired by each 3D perception camera is stitched together by the coordinate system transformation relationship of each 3D perception camera obtained by the calibration module to obtain complete global 3D spatial information of the measurement scene.

[0016] The intelligent tracking module, based on a deep learning-based target recognition and segmentation network, identifies and segments objects in real time from the global 3D spatial information output by the 3D global perception module, outputs labels for each 3D information, and outputs the 3D information of each object; it acquires pose from the real-time output of the power arm, multi-degree-of-freedom motion device, measurement unit, and driven measurement arm, and adjusts the CAD models of the power arm, multi-degree-of-freedom motion device, measurement unit, and driven measurement arm; it acquires joint data of the power arm to adjust the angles of each joint on the CAD model of the power arm, achieving a one-to-one correspondence between the joint angles on the CAD model of the power arm and the joint angles of the power arm in the measurement scene;

[0017] The 3D data alignment module aligns the labeled 3D information output by the intelligent tracking module with the coordinates of the CAD model, calculates the corresponding coordinate system transformation matrix, i.e., the alignment matrix, and outputs the matrix, the 3D spatial information aligned to the global coordinate system, and the CAD model. The inputs include the labeled 3D spatial information output by the intelligent tracking module, the CAD models of each object in the measurement scene, the features to be measured output by the measurement feature extraction module, and the CAD model of the joint angles for adjusting joint angles output by the intelligent tracking module.

[0018] The path planning module employs reinforcement learning. Its input consists of aligned 3D information and the feature to be measured, output by the 3D data alignment module. Its output is the planned measurement point. Based on the motion path of the active arm and the multi-degree-of-freedom motion device determined by the measurement point, the module intelligently plans the motion path of the active arm, enabling the measurement unit to reach the planned measurement feature to achieve the measurement of the feature.

[0019] The motion and collision avoidance detection module controls the movement of the active arm and detects collisions. It includes offline collision avoidance detection, motion execution, and online collision avoidance detection. Offline collision avoidance detection detects collisions by simulating the motion process based on the measurement points, the active arm, and the motion path of the multi-degree-of-freedom motion device output by the path planning module. Motion execution removes the measurement points and corresponding paths where collisions occurred based on the collision detection results simulated by offline collision avoidance detection, and controls the active arm to execute the remaining motion path, allowing the measurement unit to reach the remaining measurement points. Online collision avoidance detection uses the output results of the 3D global perception module, 3D data alignment module, and intelligent tracking module to detect whether there is any overlap in the spatial information of different parts of the measurement scene. If there is overlap, the measurement points and motion paths will be actively adjusted to prevent collisions.

[0020] The measurement module executes measurement tasks, and its output is the measurement result output by the measurement unit.

[0021] Furthermore, in the aforementioned scene-aware, micron-level precision dual robotic arm measuring device, the end flange of the power arm has mounting holes and is fixedly connected to a multi-degree-of-freedom motion device via threaded fasteners; the connecting component is a clamp, which is installed on the third axis of the driven measuring arm via bolt fasteners, and the clamp is provided with rope holes; the fixed pulley is installed on the environmental wall.

[0022] The present invention provides a scene-aware, micron-level precision dual-robotic arm measurement method, comprising the following steps:

[0023] 1) Calibrate the measuring equipment, calibrate the optical-based non-contact measuring device of the measuring unit, calibrate the driven measuring arm and contact measuring head, calibrate each device in the measurement scenario, and obtain the coordinate system transformation relationship between each part;

[0024] 2) Extract measurement features: Use a large language model to extract the measurement features from the input CAD model of the object to be measured. The input is the CAD model and manufacturing detail requirements, and the output is the measurement features. Obtain valid measurable elements from the CAD model to provide prior measurements, i.e., measurement targets, for subsequent measurements;

[0025] 3) Obtain complete global 3D spatial information of the measurement scene. Use multiple 3D sensing cameras to acquire 3D information in the measurement scene, including point information and corresponding color information. Then, use the coordinate system transformation relationship in step 1) to align the 3D spatial information of the measurement scene in the global coordinate system G to obtain a complete 3D representation of the measurement scene. Acquire 3D spatial information from real measurement scenarios;

[0026] 4) Identify and segment the 3D spatial information in the real measurement scene, and obtain the 3D spatial information of the object to be measured, the power arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measurement unit in the real measurement scene. The inputs are the CAD model of the object to be measured, the CAD models of the power arm and the driven measuring arm, and the 3D representation obtained in step 3). The output is the 3D information of the object under test. 3D information of the power arm and the driven measuring arm and 3D spatial information of multi-degree-of-freedom motion devices and measurement units and Employing deep learning-based object recognition and segmentation networks to identify and segment objects from 3D representations The system acquires the 3D spatial information of the object to be measured and its labels, namely the object to be measured, the power arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measuring unit.

[0027] 5) Collect and acquire real-time pose parameters of the power arm, multi-degree-of-freedom motion device and driven measuring arm, adjust their corresponding CAD models, acquire the joint data of the power arm to adjust the angle of each joint on the CAD model of the power arm, and realize a one-to-one correspondence between the joint angles on the CAD model of the power arm and the joint angles of the power arm in the measurement scene.

[0028] 6) Perform 3D data alignment, with the input being the multiple 3D spatial information output in step 4). and The output includes the CAD model from step 5); the output is an alignment matrix and aligned 3D spatial information. The CAD model of the object under test is then compared with the 3D spatial information from step 4). Align and obtain the alignment matrix. The measurement elements extracted in step 2) Mapping to a real measurement scenario, obtain Use the parameters calibrated in step 1) Measurement elements Mapped to the active arm base coordinate system, denoted as Employ iterative nearest-point methods or machine learning-based 3D data alignment models;

[0029] 7) Plan the motion path using reinforcement learning. Based on the planned motion path, the motion execution devices are the power arm and the multi-degree-of-freedom motion device. The inputs are the CAD model of the power arm, the CAD model of the multi-degree-of-freedom motion device, the CAD model of the driven arm, and the measurement elements output from step 6). The output includes the planned measurement points and the motion paths of the active arm and the motion device with degrees of freedom.

[0030] 8) Offline collision detection: Simulate the motion process based on the measurement points, power arm and multi-degree-of-freedom motion device output in step 7), and detect whether there is a collision; measurement points and paths where a collision occurs are removed.

[0031] 9) Motion and online collision detection: Based on the measurement points and motion paths output in step 8), the power arm and multi-free motion device move to the planned measurement points and detect multiple 3D spatial information output in step 4). and If there is a possibility of overlap, the measurement points and movement paths will be actively adjusted to prevent collisions, such as moving the planned measurement points a distance away from the measurement elements.

[0032] 10) Measurement: When the measurement point is reached, the measurement unit is used to measure, and the result is converted to the base coordinate system of the driven measuring arm;

[0033] 11) Repeat steps 3), 4), 5), 6), 9), and 10) above until all planned measurement points have been measured.

[0034] Furthermore, in the aforementioned scene-aware, micron-level precision dual-robotic arm measurement method, step 1) involves global calibration of the measurement equipment, establishing coordinate system transformation relationships to provide a unified coordinate system for measurement execution. This calibration calibrates multiple coordinate systems introduced by the measurement device within the measurement scene: the active arm base coordinate system K and flange coordinate system F, the driven measuring arm base coordinate system U and the coordinate system T at the contact measuring head, the multi-degree-of-freedom motion device base coordinate system D and its end-effector coordinate system M, the global coordinate system G on the global calibration object, and the coordinate systems C of each 3D sensing camera. (i) The optical-based non-contact measuring device has an inherent coordinate system S; the following parameters are obtained directly from the device: the pose of the power arm flange. The position of the measuring head is determined by the contact at the end of the driven measuring arm. Because the multi-degree-of-freedom motion device and the power arm flange are connected by fasteners, after the equipment is installed, the coordinate system F of the power arm flange overlaps with the coordinate system D of the multi-degree-of-freedom motion device, that is...

[0035] Furthermore, the above-mentioned scene-aware, micron-level precision dual-robotic arm measurement method, step 1) specifically includes the following steps:

[0036] 11) Calibration preparation: Place the global calibration object, and position the global calibration object at each ArUco feature point on the body coordinate system G. G P j and standard measurement of the center of the ball G P i The coordinates, i.e. This coordinate system serves as the global coordinate system G,o for the measurement system. G -x G y G z G Calibrate the driven measuring arm and the optical-based non-contact measuring instrument as required;

[0037] 12) Calibrate the global 3D perception system and calibrate the coordinate system C of each 3D perception camera. (i) The positional relationships are represented in the global coordinate system G; each 3D sensing camera C (i) Identify ArUco encoding and obtain the 3D feature point coordinates of ArUco encoding in the coordinate system of each 3D sensing camera.

[0038] Based on coordinates 3D coordinates in body coordinate system G P j Correspondence, calculating the transformation relationship between coordinate systems Obtain the 3D sensing camera coordinate system C (i) and C (j) Relative positional relationship

[0039]

[0040] 13) Calibrate the transformation relationship between the driven measuring arm base coordinate system and the global coordinate system, and the transformation relationship between the driven measuring arm base coordinate system U and the global coordinate system G. The coordinates of the center of a standard sphere are measured using a contact measuring head mounted at the end of the driven measuring arm. U P i And based on the measurement results in the coordinate system of the global calibration object The following relationship exists:

[0041]

[0042] 14) Calibration of the measurement unit, coordinate system transformation relationship between the optical non-contact measurement device and the contact measurement head coordinate system position T. Non-contact measurement results based on optical principles are obtained through... Transform it to the coordinate system of the measuring arm base, and output it as the final optical metrology result;

[0043] The end of the moving measuring arm is in position. At this location, the pose data is read from the measuring arm, which enables the optical-based non-contact measuring device to recognize multiple ArUco codes and obtain the 3D coordinates of the codes in coordinate system S. S P j Through this coded coordinates S P j 3D coordinates in body coordinate system G P j The correspondence between coordinate systems is calculated.

[0044]

[0045] Coordinate system pose at the contact measuring head satisfy:

[0046]

[0047] have to

[0048] 15) Determine the transformation relationship between the base coordinate system K of the boom and the global coordinate system G. Position of the multi-degree-of-freedom motion device at the end of the fixed power arm It contains three operations:

[0049] TCP calibration employs a six-point calibration method. A contact-type measuring unit at the end of the unit makes close contact with a high-precision standard sphere on the global calibration object. Coordinate system T is then used as the new TCP coordinate system for the boom, and its relationship with the boom end flange coordinate system F(D) is calculated.

[0050] Calculate the transformation relationship between the power arm base coordinate system K and the driven measuring arm base U. Keep The current position is read from both the power arm and the driven measuring arm, respectively.

[0051]

[0052] The global coordinate system transformation relationship is calculated by using the coordinate system transformation relationship between the driven boom base coordinate system U and the global coordinate system to calculate the transformation relationship between the power boom base coordinate system K and the global coordinate system G.

[0053]

[0054] 16) Calibrate the coordinate system transformation relationship between the coordinate system M of the multi-degree-of-freedom motion device and the coordinate system T of the contact measuring head. Using the acquired and Calculate

[0055] All parameters in the measurement field have been calibrated, among which and Together they determine the final pose of the measurement unit. and Measurement results applied to optical-based non-contact measuring instruments S P, so that the coordinate system of the driven measuring arm is used as the reference coordinate system for outputting measurement results. U P:

[0056]

[0057] in, It is the real-time position and orientation of the contact measuring head on the driven measuring arm at the current measurement point.

[0058] Compared with the prior art, the present invention has significant advantages and beneficial effects, specifically reflected in the following aspects:

[0059] ① This invention addresses the issue of traditional measuring arms lacking power. Its ingenious structural design allows the power arm to pull the driven measuring arm for measurement. The power arm and multi-degree-of-freedom motion device serve as the power carrier, while the driven measuring arm and measuring unit act as the measurement carrier. Fixed pulleys, counterweights, and ropes stabilize the driven measuring arm. The power arm and multi-degree-of-freedom motion device act as the active motion mechanism; their movement drives the measuring unit, which in turn drives the driven measuring arm. Measurement results are obtained through the driven measuring arm. High-precision measurement results are achieved by utilizing the high-precision characteristics of the measuring arm.

[0060] ② It enables automatic identification of measurement features and automatic path planning, effectively planning the measurement path and increasing the anti-collision capability of the measurement system, which greatly improves the measurement quality and efficiency;

[0061] ③ It enables automated measurement and features ease of use and high precision.

[0062] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing specific embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0063] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 : A schematic diagram of the structure of the device of the present invention;

[0065] Figure 2 : A schematic diagram of the structure of the measuring unit of this invention;

[0066] Figure 3 : A schematic diagram of the structure of the multi-degree-of-freedom motion device of the present invention;

[0067] Figure 4 : A schematic diagram of the architecture of the control system of this invention;

[0068] Figure 5 : A schematic diagram of the global calibration object;

[0069] Figure 6 : A schematic diagram of multiple coordinate systems introduced by the measuring equipment;

[0070] Figure 7 : A schematic diagram of the global perception module calibration;

[0071] Figure 8 : Schematic diagram of the calibration of the measuring arm and the global coordinate system;

[0072] Figure 9 A schematic diagram illustrating the transformation relationship between the coordinate system T of an optically based non-contact measuring device and a contact measuring head.

[0073] Figure 10 : Calibrate the position of the boom coordinate system K in the global coordinate system G A schematic diagram. Detailed Implementation

[0074] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0075] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, directional and ordinal terms are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0076] like Figure 1 As shown, a dual-arm measuring device with scene awareness and micron-level precision includes a power arm 1, a driven measuring arm 2, a measuring unit 3, a multi-degree-of-freedom motion device 4, multiple 3D sensing cameras 10, and a global calibration object 11 for parameter calibration. The measuring unit 3 includes an optical non-contact measuring device 31 and a contact measuring head 32 connected thereto. The end flange of the power arm 1 has mounting holes and is fixedly connected to the multi-degree-of-freedom motion device 4 by threaded fasteners. The multi-degree-of-freedom motion device 4 is connected to the non-contact measuring device 31. The end of the driven measuring arm 2 is connected to the contact measuring head 32. A connector 6 is mounted on the driven measuring arm 2. The connector 6 is a clamp. A clamp is fastened to the third axis of the driven measuring arm with bolts, and the clamp has a rope hole; a fixed pulley 7 is installed on the environmental wall, and a rope 9 passes through the fixed pulley 7, with one end connected to the connector 6 and the other end connected to the counterweight 8; 3D sensing cameras 10 are arranged at multiple locations in the measurement scene to cover the measurement environment; the power arm 1 and the multi-degree-of-freedom motion device 4 serve as active motion mechanisms, and their movement drives the measurement unit 3 to move, which in turn drives the driven measuring arm 2 connected to the measurement unit 3 to move, and the measurement results are obtained through the driven measuring arm 2; the counterweight 8 is used to stabilize the driven measuring arm 2; multiple 3D sensing cameras 10 are used for measurement environment perception; and a global calibration object 11 is used to calibrate parameters.

[0077] like Figure 2 The non-contact measuring device 31 is fixed to the end of the multi-degree-of-freedom motion device 4 by screws, and the contact measuring head 32 is connected to the non-contact measuring device 31 by clamp 33. The non-contact measuring device 31 is a structured light 3D scanner or a laser 3D scanner, and the contact measuring head 32 is a contact measuring head of an articulated arm coordinate measuring machine.

[0078] The multi-degree-of-freedom motion device 4 has two degrees of freedom: linear motion along the extension and retraction of the rod and rotational motion about the rod, such as... Figure 3The system includes a servo motor 41, a diaphragm coupling 42, a ball screw 43, a screw nut 44, a telescopic rod 45, a telescopic rod guide seat 46, a motor mounting base 47, a frameless torque motor 48, a drive shaft 49, a connecting rod 410, a linear guide rail 411, and a base 412. The servo motor 41 and the telescopic rod guide seat 46 are mounted on the base 412. The servo motor 41 is connected to the ball screw 43 through the diaphragm coupling 42. The ball screw 43 is engaged with the screw nut 44. The screw nut 44 is connected and fixed to the telescopic rod 45. The telescopic rod 45 passes through the guide hole of the telescopic rod guide seat 46 and is connected to the motor mounting base 47. The telescopic rod 45 is slidably engaged with the linear guide rail 411. The frameless torque motor 48 is mounted in the motor mounting base 47 and is connected to the connecting rod 410 through the drive shaft 49.

[0079] The servo motor 41 drives the ball screw 43 to rotate through the diaphragm coupling 42, which in turn causes the screw nut 44 to move back and forth, driving the telescopic rod 45 to move linearly along the linear slide rail 411, which in turn drives the motor mounting base 47 to move linearly. At the same time as the motor mounting base 47 moves linearly, the frameless torque motor 48 installed therein drives the rotating shaft 49 to rotate, which in turn drives the connecting rod 410 to rotate around the axis of the frameless torque motor 48, realizing the linear and rotary multi-axis motion of the connecting rod 410.

[0080] The power arm 1 is a six-degree-of-freedom robot. The control system 5 is connected to the power arm 1, the driven measuring arm 2, multiple 3D sensing cameras 10, and the multi-degree-of-freedom motion device, such as... Figure 4 The control system 5 includes a calibration module 51, a 3D global perception module 52, a measurement element extraction module 57, an intelligent tracking module 53, a 3D data alignment module 54, a path planning module 58, a motion and collision avoidance detection module 55, and a measurement module 56.

[0081] The measurement element extraction module 57 extracts measurement elements of spare parts based on a large language model, extracts information from the input CAD model of the object to be measured, and outputs the features to be measured.

[0082] Calibration module 51 calibrates and calibrates the parameters of the measuring equipment, calibrates the optical non-contact measuring device in the measuring unit 3, calibrates the driven measuring arm 2, calibrates the coordinate system transformation relationship of the equipment, i.e., the coordinate system transformation matrix, and constructs the global coordinate system;

[0083] The 3D global perception module 52 acquires spatial environment information by deploying multiple 3D perception cameras in the measurement scene. The 3D spatial information acquired by each 3D perception camera is stitched together by the coordinate system transformation relationship of each 3D perception camera acquired by the calibration module 51 to obtain complete global 3D spatial information of the measurement scene.

[0084] The intelligent tracking module 53, based on a deep learning-based target recognition and segmentation network, identifies and segments objects in real time from the global 3D spatial information output by the 3D global perception module 52, outputs labels for each 3D information, such as the object to be measured, the active arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measurement unit information, and outputs the 3D information of each object; it obtains the pose from the real-time output of the active arm, the multi-degree-of-freedom motion device, the measurement unit, and the driven measuring arm, and adjusts the CAD models of the active arm, the multi-degree-of-freedom motion device, the measurement unit, and the driven measuring arm, and obtains the joint data of the active arm to adjust the angles of each joint on the active arm CAD model, so as to achieve a one-to-one correspondence between the joint angles on the active arm CAD model and the joint angles of the active arm in the measurement scene;

[0085] The 3D data alignment module 54 aligns the labeled 3D information output by the intelligent tracking module 53 with the coordinates of the CAD model, calculates the corresponding coordinate system transformation matrix, i.e., the alignment matrix, and outputs the matrix, the 3D spatial information aligned to the global coordinate system, and the CAD model. The inputs include the labeled 3D spatial information output by the intelligent tracking module 53, the CAD models of each object in the measurement scene, the features to be measured output by the measurement feature extraction module 57, and the CAD model of the joint angles for adjusting joint angles output by the intelligent tracking module 53. It is implemented by a machine learning-based 3D data alignment model or the Iterative Closest Point (ICP) method.

[0086] The path planning module 58 adopts reinforcement learning. Its input is the aligned 3D information and the feature to be measured output by the 3D data alignment module (54). Its output is the planned measurement point. Based on the motion path of the active arm and the multi-degree-of-freedom motion device formulated according to the measurement point, the active arm motion path is intelligently planned so that the measurement unit reaches the planned measurement feature to realize the measurement of the measurement feature.

[0087] The motion and collision detection module 55 controls the movement of the active arm and detects collisions. It includes offline collision detection, motion execution, and online collision detection. Offline collision detection uses the measurement points, the active arm, and the motion path of the multi-degree-of-freedom motion device output by the path planning module 58 to simulate the motion process and detect whether there is a collision. Motion execution removes the measurement points and corresponding paths where collisions occurred based on the collision detection results simulated by offline collision detection, controls the active arm to execute the remaining motion path, and the measurement unit reaches the remaining measurement points. Online collision detection uses the output results of the 3D global perception module 52, the 3D data alignment module 54, and the intelligent tracking module 53 to detect whether there is a possibility of overlap in the spatial information of each part in the measurement scene. If there is an overlap, the measurement points and motion paths will be actively adjusted to prevent collisions.

[0088] Measurement module 56 executes measurement tasks, and its output is the measurement result output by the measurement unit.

[0089] like Figure 5 The global calibration object 11 consists of three or more calibration units, distributed in a spherical or planar form, connected by linkages to form a single unit. Each calibration unit contains three standard spheres and one ArUco code. The three non-collinear standard spheres construct a stable spatial triangular reference, constraining the spatial rotational and translational degrees of freedom of the driven measuring arm, and improving the consistency of calibration parameters under multiple postures of the contact measuring head. Global coordinate system G, standard sphere center coordinates... G P j ArUco feature points G P j Define coordinate system as "coordinate system *" and coordinate system "o". * -x * y * z * "They are in the same coordinate system."

[0090] Global calibration of measuring equipment is crucial. Establishing coordinate system transformation relationships among various devices provides a unified coordinate system for each measurement execution step, such as... Figure 6 The calibration measurement scenario includes multiple coordinate systems introduced by the measuring device: the active arm base coordinate system K and flange coordinate system F, the driven measuring arm (including the end contact measuring head) base coordinate system U and the coordinate system T at the contact measuring head, and the multi-degree-of-freedom motion device base coordinate system D and its end coordinate system M (real-time coordinate system transformation relationship). (Read from the device), global coordinate system G on the global calibration object, and coordinate systems C of each 3D sensing camera. (i) In the optical-based non-contact measuring device with its inherent coordinate system S, the following parameters do not require calibration and are obtained in real time from the system's constituent devices: the pose of the power arm flange. The position of the measuring head is determined by the contact at the end of the driven measuring arm. Because the multi-degree-of-freedom motion device and the power arm flange are connected by fasteners, after the equipment is installed, the coordinate system F of the power arm flange overlaps with the coordinate system D of the multi-degree-of-freedom motion device, that is... The accuracy of the calibration will not affect the progress of the measurement system. The accuracy of the measurement system depends on the accuracy of the driven measuring arm. The following calibration ensures that the global calibration object will not move; the specific steps are as follows:

[0091] 11) Calibration preparation: Place the global calibration object, such as... Figure 6 From the manufacturer's global calibration object, each ArUco feature point in the body coordinate system G. G P j and standard measurement of the center of the ball G P iThe coordinates, i.e. This coordinate system serves as the global coordinate system G,o for the measurement system. G -x G y G z G ; calibrate the driven measuring arm and the optical-based non-contact measuring instrument according to the manufacturer's requirements;

[0092] 12) Calibrate the global 3D perception system and calibrate the coordinate system C of each 3D perception camera. (i) The positional relationship is represented in the global coordinate system G, such as... Figure 7 Each 3D sensing camera C (i) Identify ArUco encoding and obtain the 3D feature point coordinates of ArUco encoding in the coordinate system of each 3D sensing camera.

[0093] Based on coordinates 3D coordinates in body coordinate system G P j Correspondence, calculating the transformation relationship between coordinate systems Obtain the 3D sensing camera coordinate system C (i) and C (j) Relative positional relationship

[0094]

[0095] 13) Calibrate the transformation relationship between the driven measuring arm's base coordinate system and the global coordinate system, such as... Figure 8 The transformation relationship between the base coordinate system U of the driven measuring arm and the global coordinate system G The coordinates of the center of a standard sphere are measured using a contact measuring head mounted at the end of the driven measuring arm. U P i And based on the measurement results in the coordinate system of the global calibration object The following relationship exists:

[0096]

[0097] 14) Calibration of the measurement unit, coordinate system transformation relationship between the optical non-contact measurement device and the contact measurement head coordinate system position T. like Figure 9 Optical non-contact measurement results are obtained through Transform it to the coordinate system of the measuring arm base, and output it as the final optical metrology result;

[0098] The end of the moving measuring arm is in position. At this location, the pose data is read from the measuring arm, which enables the optical-based non-contact measuring device to recognize multiple ArUco codes and obtain the 3D coordinates of the codes in coordinate system S.S P j Through this coded coordinates S P j 3D coordinates in body coordinate system G P j The correspondence between coordinate systems is calculated, and the transformation relationship between coordinate systems is determined.

[0099]

[0100] Coordinate system pose at the contact measuring head satisfy:

[0101]

[0102] have to

[0103] 15) Determine the transformation relationship between the base coordinate system K of the boom and the global coordinate system G. Position of the multi-degree-of-freedom motion device at the end of the fixed power arm like Figure 10 It contains three operations:

[0104] a. TCP calibration: A six-point calibration measurement method is used. The contact measuring head in the measuring unit makes close contact with a high-precision standard sphere on the global calibration object. The coordinate system T is then used as the new TCP coordinate system for the boom, and its relationship with the boom end flange coordinate system F(D) is calculated.

[0105] b. Calculate the transformation relationship between the power arm base coordinate system K and the driven measuring arm base U. Keep The current position is read from both the power arm and the driven measuring arm, respectively.

[0106]

[0107] c. Calculation of global coordinate system transformation relationship: Using the coordinate system transformation relationship between the driven measuring arm base U and the global coordinate system, calculate the transformation relationship between the power arm base coordinate system K and the global coordinate system G.

[0108]

[0109] 16) Calibrate the coordinate system transformation relationship between the coordinate system M of the multi-degree-of-freedom motion device and the coordinate system T of the contact measuring head. Using the acquired and Calculate

[0110] All parameters in the measurement field have been calibrated, among which and Together they determine the final pose of the measurement unit. and Measurement results applied to optical-based non-contact measuring instruments S P, so that the coordinate system of the driven measuring arm is used as the reference coordinate system for outputting measurement results. U P:

[0111]

[0112] in, It is the real-time position and orientation of the contact measuring head on the driven measuring arm at the current measurement point.

[0113] The specific steps of the scene-aware, micron-level precision dual-robotic arm measurement method are as follows:

[0114] 1) Calibrate the measuring equipment, calibrate the optical-based non-contact measuring device of measuring unit 3, calibrate the driven measuring arm 2 and the contact measuring head, calibrate each device in the measurement scenario, and obtain the coordinate system transformation relationship between each part;

[0115] 2) Extract measurement features: Use CAD feature extraction methods to extract the measurement features from the input CAD model of the object to be measured. The input is the CAD model and manufacturing detail requirements, and the output is the measurement features. For example, the QueryCAD (2025, Saira Arshad), Automatic feature recognition and dimensional attributes extraction from cad models for hybrid additive-subtractive manufacturing (2024, Muhammad Tayyab Khan) method obtains effective measurable features from CAD models, providing priors, i.e., measurement targets, for subsequent measurements;

[0116] 3) Obtain complete global 3D spatial information of the measurement scene. Use multiple 3D sensing cameras to acquire 3D information in the measurement scene, including point information and corresponding color information. Then, use the coordinate system transformation relationship in step 1) to align the 3D spatial information of the measurement scene in the global coordinate system G to obtain a complete 3D representation of the measurement scene. Acquire 3D spatial information from real measurement scenarios;

[0117] 4) Identify and segment the 3D spatial information in the real measurement scene, and obtain the 3D spatial information of the object to be measured, the power arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measurement unit in the real measurement scene. The inputs are the CAD model of the object to be measured, the CAD models of the power arm and the driven measuring arm, and the 3D representation obtained in step 3). The output is the 3D information of the object under test. 3D information of the power arm and the driven measuring arm and 3D spatial information of multi-degree-of-freedom motion devices and measurement units and Employing deep learning-based object recognition and segmentation networks to identify and segment objects from 3D representations The 3D spatial information and labels of the object to be measured are obtained, namely the object to be measured, the power arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measuring unit. For example, the DPRNet (2019, Arshad Saira) and 3d-sis (2019, Ji Hou) methods can be used.

[0118] 5) Collect and acquire real-time pose parameters of the power arm, multi-degree-of-freedom motion device and driven measuring arm, adjust their corresponding CAD models, acquire the joint data of the power arm to adjust the angle of each joint on the CAD model of the power arm, and realize a one-to-one correspondence between the joint angles on the CAD model of the power arm and the joint angles of the power arm in the measurement scene.

[0119] 6) Perform 3D data alignment, with the input being the multiple 3D spatial information output in step 4). and The output includes the CAD model from step 5); the output is an alignment matrix and aligned 3D spatial information. The CAD model of the object under test is then compared with the 3D spatial information from step 4). Align and obtain the alignment matrix. The measurement elements extracted in step 2) Mapping to a real measurement scenario, obtain Use the parameters calibrated in step 1) Measurement elements Mapped to the active arm base coordinate system, denoted as The Iterative Closest Point (ICP) method or a machine learning-based 3D data alignment model can be used, such as DeepVCP (2019, Weixin Lu), PREDATOR (2021, Shengyu Huang), and PointNetLK (2019, Yasuhiro Aoki).

[0120] 7) Plan the motion path using a learning-based method. Based on the planned motion path, the motion execution devices are the power arm and the multi-degree-of-freedom motion device. The input consists of the CAD model of the power arm, the CAD model of the multi-degree-of-freedom motion device, the CAD model of the driven arm, and the measurement elements output from step 6). The output is the planned measurement points and the motion paths of the active arm and the motion device with degrees of freedom, such as: Reinforcement learning in robotic motion planning by combined experience-based planning and self-imitation learning (2023, Sha Luo), Model-based contextual reinforcement learning for robotic cooperativemanipulation (2025, Wenrui Wang) methods;

[0121] 8) Offline collision detection: Simulate the motion process based on the measurement points, power arm, and motion path of the multi-degree-of-freedom motion device output in step 7), and detect whether there is a collision; using RobotDK and MoveIt robot simulation software, the measurement points and paths where a collision occurred are removed.

[0122] 9) Motion and online collision detection: Based on the measurement points and motion paths output in step 8), the power arm and multi-free motion device move to the planned measurement points and detect multiple 3D spatial information output in step 4). and If there is a possibility of overlap, the measurement points and movement paths will be actively adjusted to prevent collisions, such as moving the planned measurement points a distance away from the measurement elements.

[0123] 10) Measurement: When the measurement point is reached, the measurement unit is used to measure, and the result is converted to the base coordinate system of the driven measuring arm;

[0124] 11) Repeat steps 3), 4), 5), 6), 9), and 10) above until all planned measurement points have been measured.

[0125] In summary, this invention addresses the issue of traditional measuring arms lacking power. Its ingenious structural design allows the power arm to pull the driven measuring arm for measurement. The power arm and multi-degree-of-freedom motion device serve as the power carrier, while the driven measuring arm and measuring unit act as the measurement carrier. A fixed pulley, counterweight, and rope stabilize the driven measuring arm. The power arm and multi-degree-of-freedom motion device act as the active motion mechanism, their movement driving the measuring unit, which in turn drives the driven measuring arm. Measurement results are obtained through the driven measuring arm. Utilizing the high-precision characteristics of the measuring arm, high-precision measurement results are achieved.

[0126] This invention enables automatic identification of measurement features and automatic path planning, effectively planning the measurement path and increasing the anti-collision capability of the measurement system, thereby improving measurement quality and efficiency.

[0127] This invention enables automated measurement and features ease of use and high precision.

[0128] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the invention should be included within the scope of protection of the invention. It should be noted that similar reference numerals and letters in the following figures denote similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0129] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

[0130] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

Claims

1. A dual-arm robotic arm measuring device with scene awareness and micron-level precision, characterized in that: The system includes a power arm (1), a driven measuring arm (2), a measuring unit (3), a multi-degree-of-freedom motion device (4), multiple 3D sensing cameras (10), and a global calibrator (11) for calibrating parameters. The measuring unit (3) includes an optical non-contact measuring device (31) and a contact measuring head (32) connected thereto. The end of the power arm (1) is connected to the multi-degree-of-freedom motion device (4), which is connected to the non-contact measuring device (31). The end of the driven measuring arm (2) is connected to the contact measuring head (32), and a connector (6) is mounted on the driven measuring arm (2). The rope (9) passes through the fixed pulley (7), one end is connected to the connector (6), and the other end is connected to the counterweight (8); multiple 3D sensing cameras (10) are arranged in the measurement scene to cover the measurement environment; the power arm (1) and the multi-degree-of-freedom motion device (4) are active motion mechanisms, and their motion drives the measurement unit (3) to move, which in turn drives the driven measurement arm (2) connected to the measurement unit (3) to move. The measurement results are obtained through the driven measurement arm (2), the counterweight (8) is used to stabilize the driven measurement arm (2), multiple 3D sensing cameras (10) are used to measure the environment perception, and the global calibration object (11) is used to calibrate the parameters.

2. The scene-aware, micron-level precision dual-arm measuring device according to claim 1, characterized in that: The non-contact measuring device (31) is fixed to the end of the multi-degree-of-freedom motion device (4) by screws, and the contact measuring head (32) is connected to the non-contact measuring device (31) by clamp (33).

3. The scene-aware, micron-level precision dual-arm measuring device according to claim 1 or 2, characterized in that: The non-contact measuring device (31) is a structured light 3D scanner or a laser 3D scanner, and the contact measuring head (32) is a contact measuring head of an articulated arm coordinate measuring machine.

4. The scene-aware, micron-level precision dual-arm measuring device according to claim 1, characterized in that: The multi-degree-of-freedom motion device (4) includes a servo motor (41), a diaphragm coupling (42), a ball screw (43), a screw nut (44), a telescopic rod (45), a telescopic rod guide seat (46), a motor mounting base (47), a frameless torque motor (48), a drive shaft (49), a connecting rod (410), a linear slide rail (411), and a base (412). The servo motor (41) and the telescopic rod guide seat (46) are mounted on the base (412). The servo motor (41) is connected to the base via the diaphragm coupling. The device (42) is connected to the ball screw (43), the ball screw (43) is engaged with the screw nut (44), the screw nut (44) is connected and fixed to the telescopic rod (45), the telescopic rod (45) passes through the guide hole of the telescopic rod guide seat (46) and is connected to the motor mounting seat (47), and the telescopic rod (45) is slidably engaged with the linear guide rail (411); the frameless torque motor (48) is installed in the motor mounting seat (47), and the frameless torque motor (48) is connected to the connecting rod (410) through the transmission shaft (49).

5. The dual robotic arm measuring device with scene awareness and micron-level precision according to claim 1, characterized in that: The global calibration object (11) consists of three or more calibration units, which are distributed in a spherical or planar form and connected by a linkage to form a whole. Each calibration unit contains three standard spheres and one ArUco code. The three non-collinear standard spheres construct a stable spatial triangular reference, which constrains the spatial rotational and translational degrees of freedom of the driven measuring arm.

6. The scene-aware, micron-level precision dual-arm measuring device according to claim 1, characterized in that: It also includes a control system (5) connected to the power arm (1), the driven measuring arm (2), multiple 3D sensing cameras (10), and multiple free motion devices (4). The control system (5) includes a calibration module (51), a 3D global sensing module (52), a measurement element extraction module (57), an intelligent tracking module (53), a 3D data alignment module (54), a path planning module (58), a motion and collision avoidance detection module (55), and a measurement module (56). The measurement element extraction module (57) extracts the measurement elements of spare parts based on the large language model, extracts information from the CAD model of the input object to be measured, and outputs the features to be measured. The calibration module (51) calibrates and calibrates the parameters of the measuring equipment, calibrates the non-contact optical measuring device in the measuring unit (3), calibrates the driven measuring arm (2), calibrates the coordinate system transformation relationship of the equipment, i.e. the coordinate system transformation matrix, and constructs the global coordinate system. The 3D global perception module (52) acquires spatial environment information by deploying multiple 3D perception cameras in the measurement scene. The 3D spatial information acquired by each 3D perception camera is spliced ​​together by the coordinate system transformation relationship of each 3D perception camera acquired by the calibration module (51) to obtain complete global 3D spatial information of the measurement scene. The intelligent tracking module (53) is based on a deep learning target recognition and segmentation network. It identifies and segments objects in real time from the global 3D spatial information output by the 3D global perception module (52), outputs the labels of each 3D information, and outputs the 3D information of each object. It obtains the pose from the real-time output of the power arm, multi-degree-of-freedom motion device, measurement unit and driven measurement arm, and adjusts the CAD models of the power arm, multi-degree-of-freedom motion device, measurement unit and driven measurement arm. It obtains the joint data of the power arm to adjust the angle of each joint on the CAD model of the power arm, so as to realize the one-to-one correspondence between the joint angle on the CAD model of the power arm and the joint angle of the power arm in the measurement scene. The 3D data alignment module (54) aligns the tagged 3D information output by the intelligent tracking module (53) with the coordinates of the CAD model, calculates the corresponding coordinate system transformation matrix, i.e. the alignment matrix, and outputs the matrix, the 3D spatial information aligned to the global coordinate system, and the CAD model. The inputs include the tagged 3D spatial information output by the intelligent tracking module (53), the CAD models of each object in the measurement scene, the features to be measured output by the measurement element extraction module (57), and the CAD model of the joint angle for adjusting the joint angle output by the intelligent tracking module (53). The path planning module (58) adopts reinforcement learning. Its input is the aligned 3D information and the feature to be measured output by the 3D data alignment module (54). Its output is the planned measurement point. Based on the motion path of the active arm and the multi-degree-of-freedom motion device formulated according to the measurement point, the active arm motion path is intelligently planned so that the measurement unit reaches the planned measurement feature to realize the measurement of the measurement feature. The motion and collision detection module (55) controls the movement of the active arm and detects collisions. It includes offline collision detection, motion execution and online collision detection. The offline collision detection detects whether there is a collision in the simulated motion process based on the measurement points output by the path planning module (58), the motion path of the active arm and the multi-degree-of-freedom motion device. The motion execution removes the collision points and corresponding paths that have collided based on the collision detection results simulated by the offline anti-collision detection, and controls the active arm to execute the remaining motion path so that the measurement unit reaches the remaining measurement points; the online anti-collision detection uses the output results of the 3D global perception module (52), the 3D data alignment module (54) and the intelligent tracking module (53) to detect whether the spatial information of each part in the measurement scene may overlap. If there is overlap, the measurement points and motion paths will be actively adjusted to prevent collisions. The measurement module (56) performs the measurement task and its output is the measurement result output by the measurement unit.

7. The dual robotic arm measuring device with scene awareness and micron-level precision according to claim 1, characterized in that: The end flange of the power arm (1) has mounting holes and is fixedly connected to the multi-degree-of-freedom motion device (4) by threaded fasteners; the connecting piece (6) is a clamp, which is installed on the third axis of the driven measuring arm by bolt fasteners, and the clamp is provided with rope holes; the fixed pulley (7) is installed on the environmental wall.

8. A dual-arm measurement method with scene awareness and micron-level precision, characterized in that: The steps are as follows: 1) Calibrate the measuring equipment, calibrate the optical non-contact measuring device of the measuring unit (3), calibrate the driven measuring arm (2) and the contact measuring head, calibrate each device in the measuring scenario, and obtain the coordinate system transformation relationship between each part; 2) Extract measurement features: Use a large language model to extract the measurement features from the input CAD model of the object to be measured. The input is the CAD model and manufacturing detail requirements, and the output is the measurement features. Obtain valid measurable elements from the CAD model to provide prior measurements, i.e., measurement targets, for subsequent measurements; 3) Obtain complete global 3D spatial information of the measurement scene. Use multiple 3D sensing cameras to acquire 3D information in the measurement scene, including point information and corresponding color information. Then, use the coordinate system transformation relationship in step 1) to align the 3D spatial information of the measurement scene in the global coordinate system G to obtain a complete 3D representation of the measurement scene. Acquire 3D spatial information from real measurement scenarios; 4) Identify and segment the 3D spatial information in the real measurement scene, and obtain the 3D spatial information of the object to be measured, the power arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measurement unit in the real measurement scene. The inputs are the CAD model of the object to be measured, the CAD models of the power arm and the driven measuring arm, and the 3D representation obtained in step 3). The output is the 3D information of the object under test. 3D information of the power arm and the driven measuring arm and 3D spatial information of multi-degree-of-freedom motion devices and measurement units and Employing deep learning-based object recognition and segmentation networks to identify and segment objects from 3D representations The system acquires the 3D spatial information of the object to be measured and its labels, namely the object to be measured, the power arm, the driven measuring arm, the multi-degree-of-freedom motion device, and the measuring unit. 5) Collect and acquire real-time pose parameters of the power arm, multi-degree-of-freedom motion device and driven measuring arm, adjust their corresponding CAD models, acquire the joint data of the power arm to adjust the angle of each joint on the CAD model of the power arm, and realize a one-to-one correspondence between the joint angles on the CAD model of the power arm and the joint angles of the power arm in the measurement scene. 6) Perform 3D data alignment, with the input being the multiple 3D spatial information output in step 4). and The output includes the CAD model from step 5); the output is an alignment matrix and aligned 3D spatial information. The CAD model of the object under test is then compared with the 3D spatial information from step 4). Align and obtain the alignment matrix. The measurement elements extracted in step 2) Mapping to a real measurement scenario, obtain Use the parameters calibrated in step 1). Measurement elements Mapped to the active arm base coordinate system, denoted as Employ iterative nearest-point methods or machine learning-based 3D data alignment models; 7) Plan the motion path and use reinforcement learning. Based on the planned motion path, the motion execution device is a power arm and a multi-degree-of-freedom motion device. The inputs are the CAD model of the boom, the CAD model of the multi-degree-of-freedom motion device, the CAD model of the driven boom, and the measurement elements output from step 6). The output includes the planned measurement points and the motion paths of the active arm and the motion device with degrees of freedom. 8) Offline collision detection: Simulate the motion process based on the measurement points, power arm and multi-degree-of-freedom motion device output in step 7), and detect whether there is a collision; measurement points and paths where a collision occurs are removed. 9) Motion and online collision detection: Based on the measurement points and motion paths output in step 8), the power arm and multi-free motion device move to the planned measurement points and detect multiple 3D spatial information output in step 4). and If there is a possibility of overlap, the measurement points and movement paths will be actively adjusted to prevent collisions, such as moving the planned measurement points a distance away from the measurement elements. 10) Measurement: When the measurement point is reached, the measurement unit is used to measure, and the result is converted to the base coordinate system of the driven measuring arm; 11) Repeat steps 3), 4), 5), 6), 9), and 10) above until all planned measurement points have been measured.

9. The scene-aware, micron-level precision dual-robotic arm measurement method according to claim 8, characterized in that: Step 1) Global calibration of the measuring equipment: Establish coordinate system transformation relationships to provide a unified coordinate system for measurement execution. This involves calibrating multiple coordinate systems introduced by the measuring device within the measurement scenario: the active arm base coordinate system K and flange coordinate system F, the driven measuring arm base coordinate system U and the coordinate system T at the contact measuring head, the base coordinate system D of the multi-degree-of-freedom motion device and its end-effector coordinate system M, the global coordinate system G on the global calibration object, and the coordinate systems C of each 3D sensing camera. (i) The optical-based non-contact measuring device has an inherent coordinate system S; the following parameters are obtained directly from the device: the pose of the power arm flange. The position of the measuring head is determined by the contact at the end of the driven measuring arm. Because the multi-degree-of-freedom motion device and the power arm flange are connected by fasteners, after the equipment is installed, the coordinate system F of the power arm flange overlaps with the coordinate system D of the multi-degree-of-freedom motion device, that is...

10. The scene-aware, micron-level precision dual-robotic arm measurement method according to claim 9, characterized in that: Specifically, the following steps are included: 11) Calibration preparation: Place the global calibration object, and position the global calibration object at each ArUco feature point on the body coordinate system G. G P j and standard measurement of the center of the ball G P i The coordinates, i.e. This coordinate system serves as the global coordinate system G,o for the measurement system. G -x G y G z G Calibrate the driven measuring arm and the optical-based non-contact measuring instrument as required; 12) Calibrate the global 3D perception system and calibrate the coordinate system C of each 3D perception camera. (i) The positional relationships are represented in the global coordinate system G; each 3D sensing camera C (i) Identify ArUco encoding and obtain the 3D feature point coordinates of ArUco encoding in the coordinate system of each 3D sensing camera. Based on coordinates 3D coordinates in body coordinate system G P j Correspondence, calculating the transformation relationship between coordinate systems Obtain the 3D sensing camera coordinate system C (i) and C (j) Relative positional relationship 13) Calibrate the transformation relationship between the driven measuring arm base coordinate system and the global coordinate system, and the transformation relationship between the driven measuring arm base coordinate system U and the global coordinate system G. The coordinates of the center of a standard sphere are measured using a contact measuring head mounted at the end of the driven measuring arm. U P i And based on the measurement results in the coordinate system of the global calibration object The following relationship exists: 14) Calibration of the measurement unit, coordinate system transformation relationship between the optical non-contact measurement device and the contact measurement head coordinate system position T. Non-contact measurement results based on optical principles are obtained through... Transform it to the coordinate system of the measuring arm base, and output it as the final optical metrology result; The end of the moving measuring arm is in position. At this location, the pose data is read from the measuring arm, which enables the optical-based non-contact measuring device to recognize multiple ArUco codes and obtain the 3D coordinates of the codes in coordinate system S. S P j Through this coded coordinates S P j 3D coordinates in body coordinate system G P j The correspondence between coordinate systems is calculated, and the transformation relationship between coordinate systems is determined. Coordinate system pose at the contact measuring head satisfy: have to 15) Determine the transformation relationship between the base coordinate system K of the boom and the global coordinate system G. The position of the multi-degree-of-freedom motion device at the end of the fixed power arm It contains three operations: TCP calibration employs a six-point calibration method. A contact-type measuring unit at the end of the unit makes close contact with a high-precision standard sphere on the global calibration object. Coordinate system T is then used as the new TCP coordinate system for the boom, and its relationship with the boom end flange coordinate system F(D) is calculated. Calculate the transformation relationship between the power arm base coordinate system K and the driven measuring arm base U. Keep The current position is read from both the power arm and the driven measuring arm, respectively. The global coordinate system transformation relationship is calculated by using the coordinate system transformation relationship between the driven boom base coordinate system U and the global coordinate system to calculate the transformation relationship between the power boom base coordinate system K and the global coordinate system G. 16) Calibrate the coordinate system transformation relationship between the coordinate system M of the multi-degree-of-freedom motion device and the coordinate system T of the contact measuring head. Using the acquired and Calculate All parameters in the measurement field have been calibrated, among which and Together they determine the final pose of the measurement unit. and Measurement results applied to optical-based non-contact measuring instruments S P, so that the coordinate system of the driven measuring arm is used as the reference coordinate system for outputting measurement results. U P: in, It is the real-time position and orientation of the contact measuring head on the driven measuring arm at the current measurement point.

Images