Antenna coordinate system calibration method based on machine vision

Abstract

The invention relates to an antenna coordinate system calibration method based on machine vision, comprising a mechanical arm system and a camera acquisition system, the mechanical arm system comprises a mechanical arm and a mechanical arm clamping jaw, the camera acquisition system comprises a camera, and the camera is installed on the mechanical arm clamping jaw; for the measurement plane, the mechanical arm drives the mechanical arm clamping jaw to move to a preset position, a camera installed on the mechanical arm clamping jaw carries out measurement, and the current pose of the mechanical arm is combined as a primary measurement result; and performing splicing processing by combining measurement results of multiple measurements, reproducing a 3D coordinate point cloud of the target plane, and calculating target pose information through the 3D coordinate point cloud. The coordinate system is accurate in positioning; measurement results of the method can be repeatedly used; measurement results can be reproduced, the measurement process is simple and reliable, no human factors exist in the measurement process, all the measurement results are acquired by the mechanical arm and the camera, and interference is small; the measurement result of the invention is 3D visible.

Description

Technical Field

[0001] The present invention relates to the technical field related to robot three-dimensional vision, and in particular to an antenna coordinate system calibration method based on machine vision. Background Art

[0002] The antenna coordinate system calibration method based on machine vision is a technology used for antenna positioning and attitude measurement. It uses machine vision technology to capture images or videos of the antenna through a camera, and then uses image processing and computer vision algorithms to extract and analyze the characteristics of the antenna to determine its coordinates and attitude in space. This field involves multiple disciplines, including electronic engineering, computer science, optical engineering, and mathematics. In this field, researchers need to develop efficient image processing and computer vision algorithms to achieve high-precision antenna coordinate and attitude measurement. At the same time, it is also necessary to design and optimize the camera and lighting system to ensure that clear and accurate antenna images are captured. The antenna coordinate system calibration method based on machine vision has broad application prospects in antenna testing, satellite communications, aerospace, and defense. It can help engineers and technicians quickly and accurately measure and calibrate the position and attitude of the antenna, and improve the performance and reliability of the system.

[0003] The existing technology currently includes the following methods:

[0004] Laser pointer measurement, mobile robot arm measurement and visual inspection

[0005] There are usually several methods for calibrating the laser pointer coordinate system:

[0006] 1) Calibration based on reference points: This method requires at least three reference points with known coordinates to be set in the measurement area. By measuring the position of the laser pen at these reference points, the offset and rotation parameters of the coordinate system are calculated. The advantage is that it is simple and easy to use and is suitable for small measurement areas. The disadvantage is that the coordinates of the reference points need to be known, and the laser pen needs to be able to illuminate the reference points during the measurement process.

[0007] 2) Plane-based calibration: This method requires finding a plane within the measurement area, which can be a wall, ground, or other flat surface. Place the laser pointer perpendicular to the plane and measure the coordinates of the projection point of the laser beam on the plane. By comparing the distance between the projection point and the theoretical plane, the offset and rotation parameters of the coordinate system can be calculated. The advantage is that the coordinates of the reference point do not need to be known, and it is suitable for larger measurement areas. The disadvantage is that a flat surface is required as a reference, and the laser pointer needs to be able to illuminate the plane during the measurement process.

[0008] In summary, the shortcomings of the prior art are as follows:

[0009] 1. Laser pointer (i.e. laser rangefinder) can only measure distance, not posture;

[0010] 2. It cannot be used for coordinate position calibration of irregular test pieces;

[0011] 3. There are risks of human operation in the process of moving the robotic arm, which is not feasible when working at high altitude;

[0012] 4. The positioning points are obtained by visual inspection and the accuracy cannot be guaranteed.

[0013] In view of the above-mentioned defects, the designers have actively carried out research and innovation in order to create an antenna coordinate system calibration method based on machine vision to make it more valuable for industrial use. Summary of the invention

[0014] In order to solve the above technical problems, an object of the present invention is to provide an antenna coordinate system calibration method based on machine vision.

[0015] To achieve the above object, the present invention adopts the following technical solutions:

[0016] A method for calibrating an antenna coordinate system based on machine vision includes a mechanical arm system and a camera acquisition system, wherein the mechanical arm system includes a mechanical arm and a mechanical arm clamp, and the camera acquisition system includes a camera, and the camera is installed on the mechanical arm clamp;

[0017] For the measurement plane, the robot drives the robot gripper to move to the predetermined position, and the camera installed on the robot gripper performs measurement, and the measurement is combined with the current posture of the robot as a measurement result;

[0018] Among them, a measurement result includes the following:

[0019] 1) Robotic arm position<X,Y,Z,W,P,R> ,in:

[0020] 1.X, Y, Z are the current position information of the robot arm

[0021] 2.W, P, R are the angle information of the robot arm in the XYZ direction

[0022] 2) Point cloud PCD, the measured point cloud is the point cloud data acquired by the 3D camera in the camera coordinate system;

[0023] The measurement results of multiple measurements are combined and stitched together to reproduce the 3D coordinate point cloud of the target plane, and the target posture information is calculated through the 3D coordinate point cloud.

[0024] As a further improvement of the present invention, the measuring process comprises the following steps in sequence:

[0025] Step 1: 3D reconstruction:

[0026] Using surface structured light, a two-dimensional structured light pattern is projected onto the surface of the object to achieve three-dimensional contour measurement and obtain 3D reconstruction point cloud;

[0027] Step 2: Eye-in-hand calibration:

[0028] Use the EYE-IN-HAND calibration technology to convert the point cloud based on the camera coordinate system to the point cloud based on the robot arm coordinate system;

[0029] Step 3: Point cloud stitching:

[0030] Stitch together the point cloud data collected from different sites;

[0031] Step 4: 3D plane estimation:

[0032] After the above point cloud stitching, the equation of the point cloud plane is calculated through the three-dimensional plane estimation technology to identify and estimate the plane in the point cloud data.

[0033] As a further improvement of the present invention, in step 1, monocular structured light reconstruction technology is used to perform 3D reconstruction.

[0034] As a further improvement of the present invention, after the camera detects the image position in step 2, the pixel coordinates of the camera are transformed into the spatial coordinate system of the robot arm through a calibrated coordinate conversion matrix.

[0035] As a further improvement of the present invention, the three-dimensional plane estimation technology in step 4 is a method based on geometric shapes or a method based on machine learning.

[0036] As a further improvement of the present invention, the geometry-based method uses a random sampling consensus algorithm to fit the plane.

[0037] As a further improvement of the present invention, a machine learning-based method uses deep learning technology to classify and segment point cloud data.

[0038] As a further improvement of the present invention, the measurement results can be saved in a disk of a storage system, so that each measurement result can be viewed repeatedly and can be analyzed and processed by an analysis system.

[0039] As a further improvement of the present invention, each measurement result can be replayed through a replay system.

[0040] As a further improvement of the present invention, the point cloud of each measurement result, the final spliced ​​point cloud and the resulting plane equation can all be 3D visualized through OPENGL technology.

[0041] By means of the above scheme, the present invention has at least the following advantages:

[0042] The coordinate system of the present invention is precisely positioned;

[0043] The measurement results of the present invention can be reused;

[0044] The measurement results of the present invention are reproducible, the measurement process is simple and reliable, there are almost no human factors in the measurement process, and the measurement results are all collected by the robot arm and the camera, with less interference;

[0045] The measurement results of the present invention can be visualized in 3D.

[0046] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments are briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present invention and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without creative work.

[0048] Figure 1 It is a system principle schematic diagram of a method for calibrating an antenna coordinate system based on machine vision of the present invention;

[0049] Figure 2 yes Figure 1 Schematic diagram of the robot arm system;

[0050] Figure 3 It is a schematic diagram of the principle of the EYE-IN-HAND calibration technology in the present invention. DETAILED DESCRIPTION

[0051] The specific implementation of the present invention is further described in detail below in conjunction with the accompanying drawings and examples. The following examples are used to illustrate the present invention, but are not intended to limit the scope of the present invention.

[0052] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, rather than all the embodiments. The components of the embodiment of the present invention generally described and shown in the drawings here can be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiment of the present invention provided in the drawings is not intended to limit the scope of the invention claimed for protection, but merely represents the selected embodiment of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present invention.

[0053] Example

[0054] like Figure 1 to Figure 3 As shown,

[0055] A method for calibrating an antenna coordinate system based on machine vision includes a mechanical arm system and a camera acquisition system, wherein the mechanical arm system includes a mechanical arm and a mechanical arm clamp, and the camera acquisition system includes a camera, and the camera is installed on the mechanical arm clamp;

[0056] For the measurement plane, the robot drives the robot gripper to move to the predetermined position, and the camera installed on the robot gripper performs measurement, and the measurement is combined with the current posture of the robot as a measurement result;

[0057] Among them, a measurement result includes the following:

[0058] 1) Robotic arm position<X,Y,Z,W,P,R> ,in:

[0059] 1.X, Y, Z are the current position information of the robot arm

[0060] 2.W, P, R are the angle information of the robot arm in the XYZ direction

[0061] 2) Point cloud PCD, the measured point cloud is the point cloud data acquired by the 3D camera in the camera coordinate system

[0062] The measurement results of multiple measurements are combined and stitched together to reproduce the 3D coordinate point cloud of the target plane, and the target posture information is calculated through the 3D coordinate point cloud.

[0063] Preferably, the measurement process includes the following steps in sequence:

[0064] Step 1: 3D reconstruction:

[0065] Using surface structured light, a two-dimensional structured light pattern is projected onto the surface of the object to achieve three-dimensional contour measurement and obtain 3D reconstruction point cloud;

[0066] Step 2: Eye-in-hand calibration:

[0067] Use the EYE-IN-HAND calibration technology to convert the point cloud based on the camera coordinate system to the point cloud based on the robot arm coordinate system;

[0068] Step 3: Point cloud stitching:

[0069] Stitch together the point cloud data collected from different sites;

[0070] Step 4: 3D plane estimation:

[0071] After the above point cloud stitching, the equation of the point cloud plane is calculated through the three-dimensional plane estimation technology to identify and estimate the plane in the point cloud data.

[0072] Preferably, in step 1, monocular structured light reconstruction technology is used for 3D reconstruction.

[0073] Preferably, after the camera detects the image position in step 2, the pixel coordinates of the camera are transformed into the spatial coordinate system of the robotic arm through a calibrated coordinate transformation matrix.

[0074] Preferably, the three-dimensional plane estimation technology in step 4 is a method based on geometry or a method based on machine learning.

[0075] Preferably, the geometry-based method uses a random sampling consensus algorithm to fit the plane.

[0076] Preferably, the machine learning based method uses deep learning techniques to classify and segment point cloud data.

[0077] Preferably, the measurement results can be saved in a disk of a storage system, so that each measurement result can be viewed repeatedly and can be analyzed and processed by an analysis system.

[0078] Preferably, each measurement result can be played back through a playback system.

[0079] Preferably, the point cloud of each measurement result, the final spliced ​​point cloud and the resulting plane equation can be 3D visualized by OPENGL technology.

[0080] The solution of the present invention involves the intersection of multiple disciplines, including electronic engineering, computer science, optical engineering and mathematics. In this field, efforts are made to develop efficient image processing and computer vision algorithms to achieve high-precision measurement of antenna coordinates and attitude. At the same time, it is also necessary to design and optimize the camera and lighting system to ensure that clear and accurate antenna images can be captured. The goal of these works is to improve the accuracy and reliability of antenna measurements to meet the requirements of antenna performance in different application fields.

[0081] The current existing solutions have relatively rough measurement methods, many human factors in the measurement process, large random interference in the measurement results, poor visualization of the final measurement results, difficulty in result verification, and difficulty in reproducing the results.

[0082] In view of the above technical problems, the present invention proposes the following embodiments:

[0083] A method for calibrating antenna coordinate system based on machine vision:

[0084] The camera is attached to the gripper of the robot arm. For a certain part of the measurement plane, the robot arm moves to a reasonable position, and the camera obtains the measurement result. Combined with the current position of the robot arm, it is used as a measurement result. A measurement result includes the following:

[0085] 1. Robotic arm position<X,Y,Z,W,P,R> ;

[0086] 2. Point cloud PCD.

[0087] By combining the measurement results of multiple measurements and stitching the results together, the 3D coordinate point cloud of the target plane can be reproduced, and the target pose information can be calculated through the point cloud.

[0088] The following describes the technical solution during the measurement process:

[0089] 1. 3D Reconstruction

[0090] At present, the 3D reconstruction technology is mainly divided into binocular reconstruction and monocular structured light reconstruction. Binocular reconstruction technology mainly relies on the key point matching of the acquired image, which makes it difficult to form a dense point cloud. We use monocular structured light reconstruction technology. When using surface structured light, a two-dimensional structured light pattern is projected onto the surface of the object. In this way, three-dimensional contour measurement can be achieved without scanning, and the 3D reconstructed point cloud can be quickly obtained.

[0091] The point cloud collected in this stage is based on the coordinates of the camera coordinate system, and its coordinates need to be converted to the robot coordinate system. This process requires the use of the coordinate conversion matrix from the camera coordinate system to the robot coordinate system. This matrix needs to use the EYE-IN-HAND calibration technology.

[0092] 2. EYE-IN-HAND calibration technology

[0093] After the camera is bound to the robot, the point cloud based on the camera coordinate system needs to be converted to the point cloud based on the robot coordinate system. The EYE-IN-HAND calibration technology is a calibration technology that links the robot's spatial coordinate system with the camera's pixel coordinate system. In actual control, after the camera detects the image position, the camera's pixel coordinates are transformed into the robot's spatial coordinate system through the calibrated coordinate conversion matrix, which is equivalent to installing eyes.

[0094] 3. Point cloud stitching

[0095] Point cloud stitching technology refers to the technology of stitching together point cloud data collected from different sites. Through this technology, point cloud data from multiple sites can be merged to obtain more complete and accurate point cloud data, thereby improving the quality and accuracy of 3D reconstruction and preparing for subsequent point cloud pose estimation.

[0096] 4. 3D plane estimation technology

[0097] After the point cloud is stitched, the equation of the point cloud plane needs to be calculated. Point cloud plane estimation is an important technology in computer vision and 3D reconstruction, which is used to identify and estimate planes in point cloud data. This technology is often used to identify and segment planar structures in the environment. Point cloud plane estimation methods include geometry-based methods and machine learning-based methods. Geometry-based methods usually use the RANSAC (random sampling consensus) algorithm to fit the plane, while machine learning-based methods use deep learning technology to classify and segment point cloud data.

[0098] In addition, some common point cloud plane estimation algorithms include: PCA (Principal Component Analysis), RANSAC (Random Sampling Consensus), least squares method, etc.

[0099] The above specific embodiments have the following advantages:

[0100] 1. Accurate positioning of coordinate system;

[0101] 2. The measurement results can be reused:

[0102] Because the measurement results<X,Y,Z,W,P,R,PCD> The data will be saved to disk, and the results of each acquisition can be viewed and analyzed repeatedly.

[0103] 3. The measurement results can be reproduced:

[0104] For a certain measurement result<X,Y,Z,W,P,R,PCD> , the current position of the robot arm can be accurately set to<X,Y,Z,W,P,R> , and then data collection is performed. Because this system uses structured light measurement to obtain dense point clouds, the PCD results of the two measurements are relatively consistent.

[0105] The measurement process is simple and reliable, with almost no human factors. The measurement results are all collected by the robotic arm and camera, with less interference.

[0106] 4. 3D visualization of measurement results:

[0107] Each measurement result point cloud and the final stitching point cloud, as well as the resulting plane equation, can be visualized in 3D using OPENGL technology.

[0108] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly referring to the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0109] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0110] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the technical principles of the present invention, and these improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A method for calibrating an antenna coordinate system based on machine vision, comprising a mechanical arm system and a camera acquisition system, wherein the mechanical arm system comprises a mechanical arm and a mechanical arm clamp, and the camera acquisition system comprises a camera, wherein the camera is mounted on the mechanical arm clamp; Features: For the measurement plane, the robot arm drives the robot arm gripper to move to a predetermined position, and the camera installed on the robot arm gripper performs measurement, and the measurement is combined with the current posture of the robot arm as a measurement result; Among them, a measurement result includes the following: 1) Robotic arm position<X,Y,Z,W,P,R> ; 2) Point cloud PCD; The measurement results of multiple measurements are combined and stitched together to reproduce the 3D coordinate point cloud of the target plane, and the target posture information is calculated through the 3D coordinate point cloud.

2. The antenna coordinate system calibration method based on machine vision according to claim 1, characterized in that: The measuring process comprises the following steps in sequence: Step 1: 3D reconstruction: Using surface structured light, a two-dimensional structured light pattern is projected onto the surface of the object to achieve three-dimensional contour measurement and obtain 3D reconstruction point cloud; Step 2: Eye-in-hand calibration: Use the EYE-IN-HAND calibration technology to convert the point cloud based on the camera coordinate system to the point cloud based on the robot arm coordinate system; Step 3: Point cloud stitching: Stitch together the point cloud data collected from different sites; Step 4: 3D plane estimation: After the above point cloud stitching, the equation of the point cloud plane is calculated through the three-dimensional plane estimation technology to identify and estimate the plane in the point cloud data.

3. The antenna coordinate system calibration method based on machine vision according to claim 2, characterized in that: In the step 1, the monocular structured light reconstruction technology is used to perform 3D reconstruction.

4. The antenna coordinate system calibration method based on machine vision according to claim 2, characterized in that: In step 2, after the camera detects the image position, the pixel coordinates of the camera are transformed into the spatial coordinate system of the robot arm through the calibrated coordinate conversion matrix.

5. The antenna coordinate system calibration method based on machine vision according to claim 2, characterized in that: The three-dimensional plane estimation technology in step 4 is a method based on geometry or a method based on machine learning.

6. The antenna coordinate system calibration method based on machine vision according to claim 5, characterized in that: The geometry-based method uses a random sampling consensus algorithm to fit the plane.

7. The antenna coordinate system calibration method based on machine vision according to claim 5, characterized in that: The machine learning based approach uses deep learning techniques to classify and segment point cloud data.

8. The antenna coordinate system calibration method based on machine vision according to claim 1, characterized in that: The measurement results can be saved in a disk of a storage system, so that each measurement result can be viewed repeatedly and can be analyzed and processed by an analysis system.

9. The antenna coordinate system calibration method based on machine vision according to claim 1, characterized in that: Each measurement result can be played back through the playback system.

10. The antenna coordinate system calibration method based on machine vision according to claim 1, characterized in that: The point cloud of each measurement result, the final stitched point cloud and the resulting plane equation can all be visualized in 3D using OPENGL technology.

Images