A method for improving the accuracy of robot laser scanning measurement based on CMM

Abstract

The present invention discloses a method for improving the accuracy of robot laser scanning measurement based on CMM. By using a number of discrete points measured by CMM and combining with the corresponding information of the CAD model, precise positioning and alignment of the free-form surface are achieved. Then, the robot is used to drive the laser sensor to quickly scan and measure the workpiece, and the generated point cloud data is filtered and then aligned and registered with the precise point data measured by CMM to obtain the three-dimensional detection data of the workpiece with high precision. The present invention proposes a compensation method for improving the accuracy of robot laser scanning measurement, which solves the problems of low accuracy, poor stability, complex detection path planning, and conversion and processing of measurement data in industrial applications of robot laser detection. In addition, through calibration and compensation algorithms, the detection data can be digitally and intelligently processed, and its application value and market prospect are huge.

Description

Technical Field

[0001] The present invention relates to the field of measurement technology, and particularly relates to a method for improving the measurement accuracy of robot laser scanning based on CMM. Background Art

[0002] A Coordinate Measuring Machine (CMM) is a precision measurement device. Using a CMM for precision measurement is one of the most effective methods to measure and obtain workpiece size data, and it is widely used in industries such as machinery, automotive, aviation, military, and molds. The most common type of CMM is "contact detection", and its advantages are: high detection accuracy, stable performance, especially suitable for measuring high-precision parts with regular shapes, and it is an indispensable precision detection device in the current manufacturing quality detection system. However, contact detection also has its technical drawbacks. For example, the probe of contact measurement directly contacts the measured surface, which is easy to cause probe damage, and it is not easy to measure the internal profile of complex parts and soft materials. At the same time, when using contact measurement, the measured surface has a small deformation, which cannot be applied to thin-walled complex curved surfaces with high-precision requirements in fields such as aerospace. In addition, contact measurement mostly uses a point-by-point measurement method, with low measurement efficiency, slow measurement speed, and the measurement results also require probe radius compensation. In response to these defects of contact measurement, the current application and development of non-contact measurement technology are rapid, mainly aiming to solve the above problems brought by CMM contact measurement. Non-contact measurement has the advantages of fast measurement speed, high measurement efficiency, strong implementability, no deformation of the measured surface, no damage to the probe, and no need for radius compensation, becoming the development trend of complex curved surface measurement. Most CMM manufacturers at home and abroad have adopted non-contact sensors such as lasers and cameras on CMM equipment to achieve rapid detection. However, it mainly relies on the CMM's own structure, control, and data processing. Moreover, this technology field mainly relies on foreign hardware and software suppliers to provide technology upgrades and services, with poor functional openness, high prices for upgrades and transformations, and it is difficult to be compatible with the technical advantages of contact detection and non-contact detection. In addition, currently, as the detection requirements continue to increase, especially the detection based on the 3D mathematical model direction is the future development direction of precision detection, and spatial three-dimensional compensation is one of the key technologies in the detection of complex parts. With the rapid development of robot technology and high-precision sensor technology, the application fields of robots will also develop in the direction of multi-disciplinary and high-tech cross-integration, and its technology involves many fields such as electrical and electronic, mechanical manufacturing, computer, sensors, and data processing. Therefore, the CMM contact measurement can be combined with the flexible control technology of robots and the laser detection and positioning principle, so as to have a method for robot laser detection based on CMM, solving some technical bottlenecks such as low precision, poor stability, complex detection path planning, and cumbersome measurement data conversion and processing in the industrial application of robot laser detection. In addition, by using the 3D mathematical model for precision calibration and compensation, the data processing is digitalized, further improving the precision of robot laser scanning measurement.

[0003] In the prior art: Invention patent application number: 201610602245.0, invention patent name: A method for transforming a robot coordinate system into a three-dimensional measurement coordinate system. This invention patent mainly proposes a method for transforming a robot coordinate system into a three-dimensional measurement coordinate system. This method realizes the exchange of measurement data coordinates to robot coordinates through common point extraction, transformation matrix calculation and coordinate conversion steps, but does not involve position calibration and control of scanning measurement of the robot end loading detection tool. In addition, methods for improving the accuracy of laser scanning detection by robots, as well as error compensation and data fusion algorithms for CMM precision measurement data, are not involved.

[0004] Invention patent application number: 201910596496.6, invention patent name: A robot calibration method based on a three-coordinate platform. The invention mainly proposes a robot calibration method based on a three-coordinate platform, which is to set calibration points on the three-coordinate platform, use industrial cameras for auxiliary positioning, obtain calibration point position information, analyze and measure the three-dimensional error of the platform, and use linear algorithms to compensate for the geometric error of the measurement platform to complete the error compensation of the three-coordinate platform, and establish a mapping relationship between the three-coordinate platform coordinate system and the robot basic coordinate system for robot calibration teaching. However, this method is based on a three-coordinate platform built by itself, so the accuracy of the three-coordinate platform built by it is extremely high. Summary of the invention

[0005] In view of the above problems, the present invention provides a method for improving the accuracy of robot laser scanning measurement based on CMM, which utilizes the flexible control characteristics of robots and line laser sensors, combines the high accuracy and reliability data of CMM contact measurement points and the CAD theoretical model of the measured workpiece, and performs efficient and high-precision measurement of complex curved surface workpieces, thereby realizing three-dimensional measurement and accuracy analysis of complex curved surface parts. The present invention adopts the following technical solutions:

[0006] A method for improving the measurement accuracy of robot laser scanning based on CMM, characterized by comprising the following steps:

[0007] S1. Calibrate the three-dimensional coordinate probe using the high-precision calibration ball that comes with the CMM.

[0008] S2. Perform conversion calibration between the laser sensor coordinate system and the robot end coordinate system. The specific operations are as follows:

[0009] Place a standard gauge block on the CMM platform. The position of the gauge block is placed in three orientations, parallel to the X-axis, Y-axis, and Z-axis of the CMM respectively. Then, through the robot teach pendant or offline simulation software, real-time communication and control with the robot are carried out. The laser sensor at the end of the robot is controlled to move along the X, Y, and Z directions respectively. During the rough adjustment, the position of the laser sensor can be adjusted by observing the coincidence position relationship between the beam of the laser sensor and the edge of the gauge block. For precise adjustment, the scanned point cloud data needs to be imported into the computer point cloud processing software, and data filtering is performed. The point cloud data and the CAD digital model of the standard gauge block are subjected to ICP registration to obtain the transformation matrix between the point cloud and the CAD digital model. This transformation matrix is a four-dimensional transformation matrix:

[0010]

[0011] The elements starting with the M number in this 4*4 matrix are the transformation matrix data after ICP registration. Tx, Ty, and Tz are the translation data in the X, Y, and Z directions. Since the transformation matrix after ICP registration is obtained, the arrangement method of a matrix array and the calculation formula for multiplying a point or vector by the matrix to obtain new coordinates are as follows:

[0012]

[0013] When performing the transformation operation, the four components of the current point are respectively dot-multiplied with each row of the matrix, as shown below:

[0014]

[0015] Since in the transformation matrix, the values of M30, M31, and M32 in the last row are always 0, the result of w’ depends on w. When used as a coordinate point, the w parameter is 1. In this way, all geometric transformations and vector operations can be performed and transformed using the same matrix multiplication.

[0016] S3. Calibrate the laser sensor in the CMM coordinate system. The calculation formula is:

[0017]

[0018] In the formula, represents the target pose of the robot; represents the initial position of the end of the robot; R represents the robot coordinate system; E represents the current coordinate system of the end of the robot; Ed represents the target coordinate system of the end of the robot;

[0019] Compensate the calculated error directly into the coordinate system of the end of the robot.

[0020] S4. Integrate the calibrated data with professional measurement software for CMM and offline programming simulation software for industrial robots to plan the inspection path. Through the robot laser detection virtual simulation system of the CMM coordinate measuring machine, conduct virtual simulation on CMM inspection and robot laser scanning inspection to confirm whether it is safe and collision-free. If so, proceed to the next step; if not, re-plan the inspection path.

[0021] S5. The CMM measures the workpiece to be measured to obtain sparse point clouds.

[0022] S6. The robot drives the laser sensor to first scan and measure the standard ball to obtain the point cloud data of the calibration ball, and then scan and measure the workpiece to be measured along the inspection path to obtain the high-density point cloud data of the workpiece.

[0023] S7. Rapidly process the high-density point cloud data scanned and measured by the laser sensor through point cloud processing or measurement software, align and register the obtained ideal point cloud data with the sparse point cloud data measured by the CMM, rapidly reconstruct the three-dimensional model of the workpiece to be measured, and compare, analyze and evaluate the three-dimensional model with the theoretical CAD model of the workpiece to be measured, so as to obtain the evaluation of the true size and accuracy of the workpiece.

[0024] Further, the robot is installed around the CMM, and the laser sensor is installed at the end of the robot, as Figure 1

[0025] The beneficial effects of the present invention are:

[0026] 1. This invention patent combines the high-precision contact measurement principle of CMM, the flexible control technology of robots, and the laser detection and positioning principle. With the help of the CAD model of the workpiece to be measured for multi-sensor information fusion, a method for robot laser detection based on CMM is proposed, and a compensation method for improving the accuracy of robot laser scanning measurement is provided, solving some technical bottlenecks such as low accuracy, poor stability, complex inspection path planning, and cumbersome measurement data conversion and processing in the industrial application of robot laser detection. In addition, through the calibrated and compensated data processing algorithm, the detection data can be digitally and intelligently processed, and its application value and market prospect are huge. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings of the embodiments will be briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present invention and do not limit the present invention.

[0028] Figure 1 It is a schematic flow chart of the present invention;

[0029] Figure 2A schematic diagram of sparse point cloud data for CMM contact detection of the present invention;

[0030] Figure 3 This is a schematic diagram of high-density point cloud data after laser scanning of the present invention;

[0031] Figure 4 The error distribution diagram of the laser scanning point cloud which is not calibrated and calculated and compensated by the present invention;

[0032] Figure 5 The results of the implementation of the present invention;

[0033] Figure 6 This is a comparison error distribution diagram of the laser detection point cloud data after calibration and compensation of the present invention and the CAD digital model;

[0034] Figure 7 Schematic diagram of the device used in the present invention. DETAILED DESCRIPTION

[0035] In order to make the purpose, technical solution and advantages of the embodiment of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings of the embodiment of the present invention. Obviously, the described embodiment is a part of the embodiment of the present invention, not all of the embodiments. Based on the described embodiment of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0036] Unless otherwise defined, the technical terms or scientific terms used in the present disclosure shall have the usual meanings understood by persons with ordinary skills in the field to which the present disclosure belongs. The words "include" or "comprise" and the like used in the present disclosure mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0037] The present invention is further described below in conjunction with the accompanying drawings and embodiments.

[0038] like Figure 1 As shown, a method for improving the laser scanning measurement accuracy of a robot based on a CMM, in this embodiment, the CMM uses Hexagon expro, and the robot uses ur5 or auboi5, and is characterized in that it includes the following steps:

[0039] S1. Calibrate the three-coordinate probe using the high-precision calibration ball that comes with the CMM.

[0040] S2. Perform the conversion and calibration between the coordinate system of the laser sensor and the coordinate system of the robot end. The specific operation is as follows:

[0041] Place a standard gauge block on the CMM platform. The gauge block is placed in three orientations, parallel to the X-axis, Y-axis, and Z-axis of the CMM respectively. Then, through the robot teach pendant or offline simulation software, communicate and control with the robot in real time, and control the laser sensor at the robot end to move along the X, Y, and Z directions respectively. During the rough adjustment, the position of the laser sensor can be adjusted by observing the coincidence position relationship between the beam of the laser sensor and the edge of the gauge block. For precise adjustment, the scanned point cloud data needs to be imported into the computer point cloud processing software, and data filtering is performed. The point cloud data and the CAD digital model of the standard gauge block are subjected to ICP registration to obtain the transformation matrix between the point cloud and the CAD digital model. This transformation matrix is a four-dimensional transformation matrix:

[0042]

[0043] The elements starting with the M number in this 4*4 matrix are the transformation matrix data after ICP registration. Tx, Ty, and Tz are the translation data in the X, Y, and Z directions. Since the transformation matrix after ICP registration is obtained, the arrangement method of a matrix array and the calculation formula for multiplying a point or vector by a matrix to obtain a new coordinate are as follows:

[0044]

[0045] When performing the transformation operation, the four components of the current point are respectively dot-multiplied with each row of the matrix, as shown below:

[0046]

[0047] Since in the transformation matrix, the values of M30, M31, and M32 in the last row are always 0, the result of w’ depends on w. When used as a coordinate point, the w parameter is 1. In this way, all geometric transformations and vector operations can be performed and transformed using the same matrix multiplication.

[0048] According to the transformation matrix data, the positions of the robot end and the laser sensor can be precisely adjusted to make their relative positions and transformation relationships with the CMM, and the corresponding relationship between the CMM coordinate system and the robot coordinate system can be established.

[0049] S3. Calibrate the laser sensor in the CMM coordinate system. The calculation formula is:

[0050]

[0051] In the formula, represents the target pose of the robot; It represents the initial position of the robot end; R represents the robot coordinate system; E represents the current coordinate system of the robot end; Ed represents the target coordinate system of the robot end;

[0052] The calculated error is directly compensated into the coordinate system of the robot end to improve the accuracy of the laser sensor for scanning and detecting the workpiece to be measured.

[0053] S4. The calibrated data is integrated with the professional measurement software of the coordinate measuring machine and the offline programming and simulation software of the industrial robot to plan the detection path; through the robot laser detection virtual simulation system of the CMM coordinate measuring machine, the virtual simulation of CMM detection and robot laser scanning detection is carried out to confirm whether it is safe and collision-free. If so, proceed to the next step; if not, re-plan the detection path.

[0054] S5. The CMM measures the workpiece to be measured to obtain a sparse point cloud. The sparse point cloud has high precision and almost no noise points.

[0055] S6. The robot drives the laser sensor to first scan and measure the standard ball to obtain the point cloud data of the calibration ball, and then scans and measures the workpiece to be measured along the detection path to obtain the high-density point cloud data of the workpiece.

[0056] S7. The high-density point cloud is large in quantity but has many noise points. The high-density point cloud data obtained by the laser sensor scanning measurement is quickly processed through point cloud processing or measurement software. The ideal point cloud data is aligned and registered with the sparse point cloud data measured by the CMM, and the three-dimensional model of the workpiece to be measured is quickly reconstructed. By comparing, analyzing and evaluating the three-dimensional model with the theoretical CAD model of the workpiece to be measured, the true size and accuracy evaluation of the workpiece can be obtained.

[0057] This embodiment measures a blade, Figure 2 It is the sparse point cloud data of CMM contact detection, Figure 3 It is the high-density point cloud data after laser scanning. Table 1 below is the error analysis of the laser scanning data without calibration and calculation compensation.

[0058] Name Offset Lower Tolerance Upper Tolerance X Y Z dX dY dZ DL Surface Point - 1 0.000 -0.025 0.025 -43.215 -8.307 -16.215 -0.006 -0.098 0.115 0.151 Surface Point - 2 0.000 -0.025 0.025 -43.221 -8.216 -16.138 -0.006 -0.100 0.117 0.154 Surface Point - 3 0.000 -0.025 0.025 -43.020 -8.397 -16.283 -0.006 -0.099 0.116 0.153 Surface Point - 4 0.000 -0.025 0.025 -43.139 -8.416 -16.304 -0.006 -0.096 0.113 0.148 Surface Point - 5 0.000 -0.025 0.025 -43.257 -8.286 -16.199 -0.006 -0.098 0.114 0.150 Surface Point - 6 0.000 -0.025 0.025 -43.184 -8.281 -16.191 -0.006 -0.099 0.116 0.153 Surface Point - 7 0.000 -0.025 0.025 -43.069 -8.335 -16.232 -0.006 -0.100 0.117 0.154 Surface Point - 8 0.000 -0.025 0.025 -43.196 -8.356 -16.256 -0.006 -0.097 0.113 0.149 Surface Point - 9 0.000 -0.025 0.025 -43.162 -8.349 -16.249 -0.006 -0.098 0.114 0.150 Surface Point - 10 0.000 -0.025 0.025 -43.122 -8.378 -16.271 -0.006 -0.098 0.114 0.150 Surface Point - 11 0.000 -0.025 0.025 -43.135 -8.356 -16.254 -0.006 -0.098 0.115 0.151 Surface Point - 12 0.000 -0.025 0.025 -43.052 -8.436 -16.317 -0.006 -0.097 0.114 0.150 Surface Point - 13 0.000 -0.025 0.025 -43.226 -8.410 -16.304 -0.006 -0.095 0.111 0.146 Surface Point - 14 0.000 -0.025 0.025 -43.271 -8.377 -16.278 -0.006 -0.095 0.111 0.146 Surface Point - 15 0.000 -0.025 0.025 -42.077 -8.120 -15.999 -0.007 -0.113 0.132 0.174 Surface Point - 16 0.000 -0.025 0.025 -41.715 -8.366 -16.169 -0.013 -0.099 0.121 0.157 Surface Point - 17 0.000 -0.025 0.025 -41.758 -7.932 -15.819 -0.014 -0.114 0.140 0.181 Surface Point - 18 0.000 -0.025 0.025 -42.887 -7.979 -15.919 -0.006 -0.108 0.126 0.166 Surface Point - 19 0.000 -0.025 0.025 -41.765 -8.258 -16.086 -0.012 -0.103 0.127 0.164 Surface Point - 20 0.000 -0.025 0.025 -42.385 -7.927 -15.850 -0.007 -0.113 0.132 0.174 Surface Point - 21 0.000 -0.025 0.025 -42.158 -8.493 -16.315 -0.012 -0.101 0.123 0.160 Surface Point - 22 0.000 -0.025 0.025 -41.640 -8.461 -16.238 -0.012 -0.094 0.116 0.150 Surface Point - 23 0.000 -0.025 0.025 -41.960 -8.334 -16.167 -0.012 -0.103 0.127 0.164 Surface Point - 24 0.000 -0.025 0.025 -42.010 -7.977 -15.874 -0.007 -0.116 0.136 0.178 Surface Point - 25 0.000 -0.025 0.025 -41.841 -8.450 -16.250 -0.012 -0.098 0.120 0.155 Surface Point - 26 0.000 -0.025 0.025 -42.167 -8.276 -16.136 -0.006 -0.110 0.128 0.169 Surface Point - 27 0.000 -0.025 0.025 -42.724 -8.191 -16.092 -0.006 -0.106 0.124 0.163 Surface Point - 28 0.000 -0.025 0.025 -41.674 -8.621 -16.372 -0.012 -0.090 0.110 0.142 Surface Point - 29 0.000 -0.025 0.025 -41.827 -8.093 -15.957 -0.013 -0.110 0.134 0.174 Surface Point - 30 0.000 -0.025 0.025 -42.299 -8.339 -16.197 -0.006 -0.107 0.126 0.165 Surface Point - 31 0.000 .0.025 0.025 -43.066 -8.280 -16.185 -0.006 -0.101 0.118 0.156 Surface Point - 32 0.000 -0.025 0.025 -42.841 -8.427 -16.299 -0.006 -0.101 0.118 0.155 Surface Point - 33 0.000 -0.025 0.025 -42.259 -8.015 -15.918 -0.007 -0.113 0.132 0.174 Surface Point - 34 0.000 -0.025 0.025 -42.354 -8.081 -15.980 -0.006 -0.111 0.130 0.171 Surface Point - 35 0.000 -0.025 0.025 -41.666 -8.427 -16.213 -0.013 -0.096 0.118 0.153 Surface Point - 36 0.000 -0.025 0.025 -42.949 -7.845 -15.808 -0.006 -0.109 0.127 0.168 Surface Point - 37 0.000 -0.025 0.025 -42.594 -7.915 -15.850 -0.007 -0.111 0.130 0.171 Surface Point - 38 0.000 -0.025 0.025 -42.515 -8.232 -16.116 -0.006 -0.107 0.125 0.165

[0059] Figure 4 It is the error distribution diagram of the laser scanning point cloud without calibration calculation compensation.

[0060] The results obtained through steps S1 - S7 of the present invention are as Figure 5 described. The error analysis of the laser detection point cloud data after calibration compensation compared with the CAD digital model is shown in Table 2. The error distribution of the laser detection point cloud data after calibration compensation compared with the CAD digital model is as Figure 6 shown.

[0061] Name Offset Lower Tolerance Upper Tolerance X Y Z dX dY dZ DL Surface Point - 1 0.000 -0.025 0.025 -4.019 -106.086 2.837 0.001 -0.000 -0.002 0.002 Surface Point - 2 0.000 -0.025 0.025 -4.232 -105.058 2.666 0.000 -0.000 -0.001 0.001 Surface Point - 3 0.000 -0.025 0.025 -0.657 -104.862 4.311 -0.000 0.000 0.000 -0.000 Surface Point - 4 0.000 -0.025 0.025 -2.704 -104.944 3.361 -0.002 0.000 0.004 -0.004 Surface Point - 5 0.000 -0.025 0.025 1.801 -104.864 5.486 -0.002 0.000 0.004 -0.005 Surface Point - 6 0.000 -0.025 0.025 3.244 -105.605 6.234 0.000 -0.000 -0.001 0.001 Surface Point - 7 0.000 -0.025 0.025 3.077 -104.596 6.097 -0.001 0.000 0.001 -0.002 Surface Point - 8 0.000 -0.025 0.025 5.363 -104.464 7.236 -0.002 0.000 0.005 -0.005 Surface Point - 9 0.000 -0.025 0.025 6.922 -104.754 8.046 0.001 -0.000 -0.001 0.001 Surface Point - 10 0.000 -0.025 0.025 8.673 -104.131 8.961 -0.001 0.000 0.001 -0.002 Surface Point - 11 0.000 -0.025 0.025 -4.077 -103.815 2.653 0.002 -0.000 -0.003 0.004 Surface Point - 12 0.000 -0.025 0.025 -1.893 -103.627 3.668 -0.001 0.000 0.002 -0.002 Surface Point - 13 0.000 -0.025 0.025 0.331 -103.448 4.714 -0.004 0.000 0.008 -0.009 Surface Point - 14 0.000 -0.025 0.025 1.557 -103.681 5.312 -0.006 0.000 0.012 -0.014 Surface Point - 15 0.000 -0.025 0.025 2.959 -103.389 5.994 -0.003 0.000 0.005 -0.006 Surface Point - 16 0.000 -0.025 0.025 5.380 -103.195 7.216 -0.005 0.000 0.009 -0.010 Surface Point - 17 0.000 -0.025 0.025 6.955 -103.185 8.041 -0.003 -0.000 0.006 -0.007 Surface Point - 18 0.000 -0.025 0.025 8.560 -102.960 8.911 -0.003 -0.000 0.005 -0.006 Surface Point - 19 0.000 -0.025 0.025 9.931 -103.023 9.669 -0.004 -0.000 0.007 -0.008 Surface Point - 20 0.000 -0.025 0.025 14.471 -103.592 12.212 -0.001 -0.000 0.001 -0.002 Surface Point - 21 0.000 -0.025 0.025 11.470 -102.961 10.541 -0.004 -0.000 0.008 -0.009 Surface Point - 22 0.000 -0.025 0.025 13.255 -102.798 11.580 0.000 0.000 -0.000 0.000 Surface Point - 23 0.000 -0.025 0.025 -3.700 -102.715 2.759 -0.001 0.000 0.003 -0.003 Surface Point - 24 0.000 -0.025 0.025 -2.749 -101.377 3.135 0.000 -0.000 -0.001 0.001 Surface Point - 25 0.000 -0.025 0.025 -1.901 -102.480 3.609 0.003 -0.000 -0.006 0.007 Surface Point - 26 0.000 -0.025 0.025 14.353 -102.276 12.281 -0.000 -0.000 0.001 -0.001 Surface Point - 27 0.000 -0.025 0.025 -1.306 -101.406 3.843 0.001 -0.000 -0.002 0.002 Surface Point - 28 0.000 -0.025 0.025 0.302 -101.446 4.626 -0.002 0.000 0.004 -0.004 Surface Point - 29 0.000 -0.025 0.025 1.485 -102.421 5.238 -0.001 0.000 0.003 -0.003 Surface Point - 30 0.000 -0.025 0.025 17.977 -102.950 14.339 -0.001 -0.000 0.001 -0.001 Surface Point - 31 0.000 -0.025 0.025 1.874 -101.049 5.390 -0.002 0.000 0.004 -0.004 Surface Point - 32 0.000 -0.025 0.025 3.059 -101.589 6.003 -0.003 0.000 0.005 -0.006 Surface Point - 33 0.000 -0.025 0.025 4.810 -101.795 6.906 -0.001 0.000 0.002 -0.002 Surface Point - 34 0.000 -0.025 0.025 4.213 -102.955 6.616 -0.005 0.000 0.010 -0.011 Surface Point - 35 0.000 -0.025 0.025 15.556 -102.409 12.981 -0.002 -0.001 0.004 -0.004

[0062] Through the above methods and experimental data analysis, the method for improving the accuracy of robot laser scanning measurement based on CMM has a very obvious effect on improving the accuracy of robot laser detection. After the point cloud data obtained by the robot loaded with a laser sensor for scanning measurement is fused with the measurement point cloud data of CMM, its detection accuracy can fully meet the set tolerance requirements.

[0063] As Figure 7 shown, the robot is installed around the CMM, and the laser sensor is installed at the end of the robot.

[0064] The above are only the preferred embodiments of the present invention, and do not impose any formal restrictions on the present invention. Although the present invention has been disclosed above with the preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some changes or modifications to the above-disclosed technical content to obtain equivalent embodiments with equivalent changes within the scope of the technical solution of the present invention. However, as long as it does not depart from the technical solution of the present invention, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A method for improving the accuracy of robot laser scanning measurement based on CMM, characterized in that, The following steps are involved: S1. Calibrate the three-dimensional coordinate probe using the high-precision calibration ball provided by the CMM. S2, perform conversion calibration between the laser sensor coordinate system and the robot end coordinate system; S3. Calibrate the laser sensor in the CMM coordinate system. The calculation expression is: Wherein, represents the target pose of the robot; represents the initial position of the robot end; R represents the robot coordinate system; E represents the current coordinate system of the robot end; E d represents the target coordinate system of the robot end; Compensate the calculated error directly into the robot end coordinate system; S4. The calibrated data is integrated with the three-coordinate professional measurement software and the industrial robot offline programming simulation software to plan the detection path; through the CMM three-coordinate measuring machine's robot laser detection virtual simulation system, the CMM detection and robot laser scanning detection are virtually simulated to confirm whether it is safe and collision-free. If so, proceed to the next step; if not, re-plan the detection path; S5, CMM measures the workpiece to obtain a sparse point cloud; S6, the robot drives the laser sensor to scan and measure the calibration ball first, obtain the point cloud data of the calibration ball, and scan and measure the workpiece to be measured along the detection path to obtain high-density point cloud data of the workpiece; S7. The high-density point cloud data scanned and measured by the laser sensor is quickly processed through point cloud processing or measurement software, and the ideal point cloud data is aligned with the sparse point cloud data measured by the CMM to quickly reconstruct the three-dimensional model of the target workpiece to be measured. The three-dimensional model is compared, analyzed and evaluated with the theoretical CAD model of the workpiece to obtain the true size and accuracy of the workpiece.

2. A method for improving the accuracy of robot laser scanning measurement based on CMM according to claim 1, characterized in that, The robot is installed around the CMM, and the laser sensor is installed at the end of the robot.

3. A method for improving the accuracy of robot laser scanning measurement based on CMM according to claim 1, characterized in that, The specific operation of conversion calibration between the laser sensor coordinate system and the robot end coordinate system in step S2 is: Place the standard gauge block on the CMM platform, and place the gauge block in three directions, parallel to the X-axis, Y-axis and Z-axis of the CMM respectively. Then, communicate and control the robot in real time through the robot teaching pendant or offline simulation software, and control the laser sensor at the end of the robot to move along the X, Y and Z directions respectively. For rough adjustment, the position of the laser sensor can be adjusted by observing the coincidence position relationship between the laser sensor beam and the edge of the gauge block. For precise adjustment, the scanned point cloud data should be imported into the computer point cloud processing software, and the data should be filtered. The point cloud data and the CAD digital model of the standard gauge block are ICP-registered to obtain the transformation matrix of the point cloud and the CAD digital model. The transformation matrix is ​​a four-dimensional transformation matrix: The elements starting with M in the 4*4 matrix are the transformation matrix data after ICP registration, and Tx, Ty, and Tz are the translation data in the X, Y, and Z directions. Since the transformation matrix after ICP registration is obtained, the arrangement of a matrix array and the calculation formula for multiplying a point or vector with the matrix to obtain the new coordinates are as follows: When performing the transformation operation, the four components of the current point are multiplied by each row of the matrix, as shown below: Since in the transformation matrix, the values of M30, M31, and M32 in the last row are always 0, the result of w’ depends on w. When used as a coordinate point, the w parameter is 1, so that all geometric transformations and vector operations can be performed and transformed using the same matrix product.

Images