CN118967816A - 5G antenna posture estimation method, device and medium based on machine vision - Google Patents
5G antenna posture estimation method, device and medium based on machine vision Download PDFInfo
- Publication number
- CN118967816A CN118967816A CN202411113530.7A CN202411113530A CN118967816A CN 118967816 A CN118967816 A CN 118967816A CN 202411113530 A CN202411113530 A CN 202411113530A CN 118967816 A CN118967816 A CN 118967816A
- Authority
- CN
- China
- Prior art keywords
- antenna
- coordinate system
- coordinates
- formula
- vector
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/70—Determining position or orientation of objects or cameras
- G06T7/73—Determining position or orientation of objects or cameras using feature-based methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/80—Analysis of captured images to determine intrinsic or extrinsic camera parameters, i.e. camera calibration
Landscapes
- Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Length Measuring Devices By Optical Means (AREA)
Abstract
本发明公开了一种基于机器视觉的5G天线位姿估计方法、设备和介质，该方法包括：1对图像进行天线检测，提取ROI区域图像；2对ROI区域天线图像进行特征提取和特征匹配；3计算特征点在原图像中的像素坐标；4基于三角测量原理计算5G天线特征点在机械臂基座坐标系下的坐标初始值；5利用LM算法，以重定位误差为目标函数，求解特征点最优坐标；6利用SVD分解法拟合最优坐标集得到天线平面的法向量；7建立天线坐标系，并计算x、y、z轴方向向量；8计算天线旋转矩阵和平移向量，得到天线位姿。本发明能提高天线位姿估计的准确性和效率，并能降低系统复杂度。
The present invention discloses a 5G antenna posture estimation method, device and medium based on machine vision, the method comprising: 1. performing antenna detection on an image and extracting an ROI area image; 2. performing feature extraction and feature matching on the ROI area antenna image; 3. calculating the pixel coordinates of the feature points in the original image; 4. calculating the initial coordinate values of the 5G antenna feature points in the mechanical arm base coordinate system based on the triangulation principle; 5. using the LM algorithm, taking the repositioning error as the objective function, solving the optimal coordinates of the feature points; 6. using the SVD decomposition method to fit the optimal coordinate set to obtain the normal vector of the antenna plane; 7. establishing an antenna coordinate system, and calculating the x, y, and z axis direction vectors; 8. calculating the antenna rotation matrix and translation vector to obtain the antenna posture. The present invention can improve the accuracy and efficiency of antenna posture estimation, and can reduce the system complexity.
Description
Technical Field
The invention relates to the field of positioning detection, in particular to a 5G antenna pose estimation method, equipment and medium based on machine vision.
Background
With the rapid development of 5G communication technology, testing of 5G antenna performance becomes particularly critical. In the 5G antenna near field detection system based on the mechanical arm, the size and the height of a sampling surface are required to be set according to the working frequency and the size of the 5G antenna, the position of the sampling surface is determined according to the pose of the 5G antenna, and the position of the sampling surface determines the sampling track of the magnetic field probe at the tail end of the mechanical arm. Only when the position and direction of the antenna are matched with the position and direction of the magnetic field probe, the electromagnetic performance of the antenna can be accurately measured.
Machine vision is capable of performing tasks such as object recognition, positioning, and measurement without contact, and plays a vital role in the development of industrial automation and robotics today. The method based on monocular vision is simple and feasible, but cannot provide real scale information of a target object; although the binocular vision-based method can recover the three-dimensional information of the object, the method has higher requirements on ambient illumination and texture, and the stereo matching algorithm is complex; although the method based on the multi-view vision can provide more comprehensive visual angle information, the method has the advantages of large operation amount, high cost and increased system complexity, and is not beneficial to wide application in actual production environments. Therefore, the above methods have certain limitations in practical applications.
Disclosure of Invention
Aiming at the defects of the prior art, the invention provides a 5G antenna pose estimation method, equipment and medium based on machine vision, so as to improve the accuracy and efficiency of antenna pose estimation and reduce the complexity of a system.
In order to achieve the above purpose, the present invention adopts the following technical scheme:
The invention discloses a 5G antenna pose estimation method based on machine vision, which is characterized by comprising the following steps of:
s1: when the monocular camera reaches the ith acquisition point In the process, the monocular camera is utilized to position the ith mechanical armAcquisition of ith antenna imagePair of YOLOv modelsThe antenna in (a) is detected to obtain an antenna image of the ith ROI areaAnd willAt the position ofThe upper left corner of the middle is marked as；
S2: antenna image for ith-1 th and ith ROI areaAndFeature extraction and feature matching are carried out to obtain non-collinear 5G antenna feature pointsAt the position ofAndThe pixel point pairs corresponding to the above、; Wherein, Representing the characteristic point of the jth 5G antenna; Representation of At the position ofThe coordinates of the two points are set up in the same plane,Representation ofAt the position ofCoordinates on; m represents the total number of feature points, and m is more than or equal to 3; And Representing a j-th pixel point pair;
S3: calculating m pixel point pairs respectively at the following positions by using the formula (1) Pixel coordinates in (a)、Wherein, the method comprises the steps of, wherein,Representation ofAt the position ofThe coordinates of the two points are set up in the same plane,Representation ofAt the position ofCoordinates of:
(1)
In the formula (1), the components are as follows, Representation ofIn the i-1 th antenna imageUpper left corner coordinates of (a);
s4: based on the triangulation principle, the jth 5G antenna characteristic point is calculated by using the formula (2) Coordinates in the robot arm base coordinate system：
(2)
In the formula (2), the amino acid sequence of the compound,An internal reference matrix of the monocular camera is obtained through camera calibration;、 The camera coordinate system is a rotation matrix and a translation vector relative to the mechanical arm base coordinate system when the monocular camera shoots for the ith time respectively, and the rotation matrix and the translation vector are obtained through coordinate transformation; 、 The camera coordinate system is relative to the rotation matrix and the translation vector of the mechanical arm base coordinate system when the monocular camera shoots for the ith-1 time respectively, and the rotation matrix and the translation vector are obtained through coordinate transformation; indicating that the monocular camera is at the first Acquisition point in secondary photographingTo the point ofDepth values of (2); indicating that the monocular camera is at the first Acquisition point in secondary photographingTo the point ofDepth values of (2);
S5: performing iterative optimization on the objective function shown in the formula (3) by using an LM algorithm, taking the calculation result of S4 as the initial value of the LM algorithm, and solving the characteristic point of the jth 5G antenna Optimal coordinates in the robot arm base coordinate system；
(3)
In the formula (3), the amino acid sequence of the compound,Representation ofAndRepositioning error between;
s6: fitting an optimal coordinate set using SVD decomposition Obtaining the normal vector of the antenna plane; Wherein T represents a transpose;
s7: with direction of normal vector of antenna plane as antenna coordinate system The axial direction vector is expressed as; Selecting a certain vector direction on an antenna plane as an antenna coordinate systemThe axial direction vector is expressed as; By calculation ofAxial direction vectorCross-product of axial direction vectorsThe axial direction vector is expressed as；
S8: obtaining a transformation matrix of the antenna coordinate system relative to the mechanical arm base coordinate system by using the method (4)NamelyAndPose of the middle antenna:
(4)
In the formula (4), the amino acid sequence of the compound, Is a rotation matrix of the antenna coordinate system relative to the robot base coordinate system,Is a translation vector of the antenna coordinate system relative to the robot base coordinate system.
The 5G antenna pose estimation method based on machine vision is also characterized in that the repositioning error in S5The method comprises the following steps:
a: calculation using (5) AndActual distance between：
(5)
B: according to the cosine law, calculate using equation (6)AndDistance of repositioning between：
(6)
C: according to the actual distanceDistance of repositioningObtaining a repositioning error of formula (7)：
(7)。
S8 is to calculate a rotation matrix using equation (8)：
(8)
S8, selecting a fixed point on the 5G antenna as an origin of an antenna coordinate system, and marking a sitting of the origin under a mechanical arm base coordinate system as a translation vector。
The invention provides an electronic device, which is characterized by comprising: a memory for storing a computer program; and the processor is used for realizing the 5G antenna pose estimation method based on machine vision when executing the computer program.
The computer readable storage medium is characterized in that the computer readable storage medium is stored with a computer program, and the computer program realizes the 5G antenna pose estimation method based on machine vision when being executed by a processor.
Compared with the prior art, the invention has the following advantages:
1. the invention adopts a mode of combining the mechanical arm and the triangulation method, and utilizes the mechanical arm to take images under different visual angles with the monocular camera, thereby effectively replacing the traditional binocular or multi-view camera system, simplifying the structure of the system, reducing the operation amount and simultaneously reducing the hardware cost.
2. The method is simple and convenient to implement, and can calculate the pose of the 5G antenna to be detected while the mechanical arm carries the probe to execute the task of detecting the radiation performance of the 5G antenna, optimize the calculated pose result based on the pose information of the mechanical arm, and improve the accuracy of pose estimation.
3. According to the method, the camera image and the pose information of the tail end of the mechanical arm are analyzed, so that the pose of the 5G antenna to be detected is estimated, priori pose information is not needed to be relied on, and the dependence of the system on priori knowledge is reduced.
Drawings
FIG. 1 is a schematic diagram of the overall structure of a system according to the present invention;
FIG. 2 is a flow chart of a pose estimation method of the present invention;
FIG. 3 is a schematic diagram of a triangulation method according to an embodiment of the present invention;
Reference numerals in the drawings: 1a monocular camera; 2, a mechanical arm; 3, a mechanical arm base; 4, a man-machine interaction unit; 5a control unit; 6, 5G antenna to be measured; and 7, carrying the object stage.
Detailed Description
In this embodiment, as shown in fig. 1, a 5G antenna pose estimation device based on machine vision includes: the device comprises a monocular camera 1, a mechanical arm 2, a mechanical arm base 3, a man-machine interaction unit 4, a control unit 5, a 5G antenna 6 to be tested and a stage 7. Wherein, the monocular camera 1 is arranged at the tail end of the mechanical arm 2 in a mode that eyes are on hands, and the optical axis direction of the monocular camera 1 is the same as the Z-axis direction of the tail end of the mechanical arm 2; the mechanical arm 2 carries the monocular camera 1 to move to an image acquisition position so as to execute an image acquisition task; the man-machine interaction unit 4 is used for monitoring the running state of the system; the control unit 5 is used for sending instructions and processing data; the 5G antenna 6 to be tested is placed on the stage 7. In this embodiment, the mechanical arm is a 6-degree-of-freedom mechanical arm, and the 5G antenna to be tested is a rectangular PCB board antenna.
In this embodiment, as shown in fig. 2, a 5G antenna pose estimation method based on machine vision includes the following steps:
s1: when the monocular camera reaches the ith acquisition point In the process, the monocular camera is utilized to position the ith mechanical armAcquisition of ith antenna imagePair of YOLOv modelsThe antenna in (a) is detected to obtain an antenna image of the ith ROI areaAnd willAt the position ofThe upper left corner of the middle is marked as；
In the present embodiment, before S1 is performed, the following preparation work needs to be completed, including:
(A) Training an antenna detection model, wherein the process is as follows: by means of the device, the mechanical arm is controlled to carry the monocular camera to collect 60 5G antenna images at different visual angles, and the labelme tool is used for marking the antenna area of each image in the data set. And randomly selecting 48 antenna images in the data set as a training set, and taking the rest 12 antenna images as a test set. The images of the training set and their corresponding labels are input YOLOv to a target detection algorithm for training. After training, the model is applied to a test set, parameter adjustment and optimization are carried out according to the test result, and a 5G antenna detection model is obtained.
(B) Calibrating the monocular camera by using a checkerboard calibration method to obtain a monocular camera internal reference matrix shown as a formula (1)The expression is:
(1)
In the formula (1), the components are as follows, AndThe normalized focal lengths on the u-axis and v-axis of the camera image coordinate system are respectively,As the focal length of the camera,AndRespectively representing the size of unit pixels on a camera image coordinate system, wherein the unit is，AndRepresenting the optical center of the camera, i.e. the intersection of the camera optical axis and the image plane.
(C) Performing hand-eye calibration on the mechanical arm-camera system by using a Tsai method to obtain a hand-eye matrix shown as (2)：
(2)
Wherein, 、Respectively represent shooting timesAnd (3) when an image is formed, a rotation matrix and a translation vector of the camera coordinate system relative to the tail end coordinate system of the mechanical arm are formed.
S2: antenna image for ith-1 th and ith ROI areaAndAnd carrying out feature extraction and feature matching. Specifically, toAndGraying, gaussian filtering and binarization are carried out, corner points and surface welding spots of an antenna are selected to serve as antenna characteristic points, a Shi-Tomas method is adopted to carry out corner point detection on an ROI image, SIFT descriptors of all the corner points are calculated, and a Brute-Force method is adopted to match the characteristic points to obtain non-collinear 5G antenna characteristic pointsAt the position ofAndThe pixel point pairs corresponding to the above、; Wherein, Representing the characteristic point of the jth 5G antenna; Representation of At the position ofThe coordinates of the two points are set up in the same plane,Representation ofAt the position ofCoordinates on; m represents the total number of feature points, and m is more than or equal to 3; And Representing a j-th pixel point pair;
s3: calculating m pixel point pairs respectively at the following positions by using the method (3) Pixel coordinates in (a)、Wherein, the method comprises the steps of, wherein,Representation ofAt the position ofThe coordinates of the two points are set up in the same plane,Representation ofAt the position ofCoordinates of:
(3)
In the formula (3), the amino acid sequence of the compound, Representation ofIn the i-1 th antenna imageUpper left corner coordinates of (a);
S4: as shown in fig. 3, based on the principle of triangulation, the jth 5G antenna characteristic point is calculated using equation (4) Coordinates in the robot arm base coordinate system：
(4)
In the formula (4), the amino acid sequence of the compound,、The camera coordinate system is a rotation matrix and a translation vector relative to the mechanical arm base coordinate system when the monocular camera shoots for the ith time respectively, and the rotation matrix and the translation vector are obtained through coordinate transformation;、 The camera coordinate system is relative to the rotation matrix and the translation vector of the mechanical arm base coordinate system when the monocular camera shoots for the ith-1 time respectively, and the rotation matrix and the translation vector are obtained through coordinate transformation; indicating that the monocular camera is at the first Acquisition point in secondary photographingTo the point ofDepth values of (2); indicating that the monocular camera is at the first Acquisition point in secondary photographingTo the point ofDepth values of (2);
In the present embodiment, the rotation matrix of the camera coordinate system relative to the robot base coordinate system is calculated by using (5) Translation vector：
(5)
In the formula (5), the amino acid sequence of the compound,Representing a transformation matrix of the camera coordinate system relative to the robot base coordinate system at the ith photograph.
S5: performing iterative optimization on the objective function shown in the formula (6) by using the LM algorithm, taking the calculation result of S4 as the initial value of the LM algorithm, and solving the jth 5G antenna characteristic pointOptimal coordinates in the robot arm base coordinate system；
(6)
In the formula (6), the amino acid sequence of the compound,Representation ofAndThe repositioning error between the two is calculated by the formula (7):
(7)
In this embodiment, the maximum iteration number of the LM optimization algorithm is set to 500, and the step size is set to . When the algorithm is executed until the maximum iteration number or step length is smaller than the given precision, the optimization process is considered to reach an optimal state, iteration is stopped, and the optimal result is used as the optimal coordinate of the jth 5G antenna characteristic point under the mechanical arm base coordinate system.
S6: fitting an optimal coordinate set using SVD decompositionObtaining the normal vector of the antenna plane; Wherein T represents a transpose;
s7: in this embodiment, the 5G antenna to be tested is a rectangular PCB board antenna, and when an antenna coordinate system is constructed, the direction of the normal vector of the antenna plane is used as the antenna coordinate system The axial direction vector is expressed as; Selecting optimal coordinates of upper left corner point on antenna planeOptimal coordinates pointing to the lower left corner pointIs the direction of the antenna coordinate systemThe axial direction vector is expressed as; By calculation ofAxial direction vectorCross-product of axial direction vectorsThe axial direction vector is expressed as；
S8: calculating a rotation matrix of the antenna coordinate system relative to the robot base coordinate system using (8)：
(8)
Selecting the optimal coordinates of the upper left corner point on the 5G antenna to be testedThe origin of the antenna coordinate system is marked as the translation vector of the antenna coordinate system relative to the mechanical arm base coordinate system by the coordinates of the origin under the mechanical arm base coordinate system. Obtaining a transformation matrix of the antenna coordinate system relative to the mechanical arm base coordinate system by using the method (9)NamelyAndPose of the middle antenna:
(9)
In this embodiment, an electronic device includes: a memory for storing a computer program for executing the above method; and a processor for executing the program stored in the memory.
In this embodiment, a computer readable storage medium stores a computer program that when executed by a processor implements the above method.
Claims (6)
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411113530.7A CN118967816B (en) | 2024-08-14 | 2024-08-14 | Machine vision-based methods, equipment, and media for 5G antenna pose estimation. |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202411113530.7A CN118967816B (en) | 2024-08-14 | 2024-08-14 | Machine vision-based methods, equipment, and media for 5G antenna pose estimation. |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN118967816A true CN118967816A (en) | 2024-11-15 |
| CN118967816B CN118967816B (en) | 2025-12-12 |
Family
ID=93390825
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202411113530.7A Active CN118967816B (en) | 2024-08-14 | 2024-08-14 | Machine vision-based methods, equipment, and media for 5G antenna pose estimation. |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN118967816B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103759716A (en) * | 2014-01-14 | 2014-04-30 | 清华大学 | Dynamic target position and attitude measurement method based on monocular vision at tail end of mechanical arm |
| CN107121125A (en) * | 2017-06-12 | 2017-09-01 | 哈尔滨工业大学 | A kind of communication base station antenna pose automatic detection device and method |
| US20190371003A1 (en) * | 2018-05-30 | 2019-12-05 | Baidu Online Network Technology (Beijing) Co., Ltd . | Monocular vision tracking method, apparatus and non-volatile computer-readable storage medium |
| WO2021217976A1 (en) * | 2020-04-28 | 2021-11-04 | 平安科技（深圳）有限公司 | Method and apparatus for controlling mechanical arm on basis of monocular visual positioning |
| CN116560395A (en) * | 2023-04-12 | 2023-08-08 | 杭州反重力智能科技有限公司 | A method for automatic inspection of communication base stations by unmanned aerial vehicle |
| CN117808880A (en) * | 2024-01-04 | 2024-04-02 | 桂林电子科技大学 | A monocular vision spatial cooperative target pose measurement method |
-
2024
- 2024-08-14 CN CN202411113530.7A patent/CN118967816B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103759716A (en) * | 2014-01-14 | 2014-04-30 | 清华大学 | Dynamic target position and attitude measurement method based on monocular vision at tail end of mechanical arm |
| CN107121125A (en) * | 2017-06-12 | 2017-09-01 | 哈尔滨工业大学 | A kind of communication base station antenna pose automatic detection device and method |
| US20190371003A1 (en) * | 2018-05-30 | 2019-12-05 | Baidu Online Network Technology (Beijing) Co., Ltd . | Monocular vision tracking method, apparatus and non-volatile computer-readable storage medium |
| WO2021217976A1 (en) * | 2020-04-28 | 2021-11-04 | 平安科技（深圳）有限公司 | Method and apparatus for controlling mechanical arm on basis of monocular visual positioning |
| CN116560395A (en) * | 2023-04-12 | 2023-08-08 | 杭州反重力智能科技有限公司 | A method for automatic inspection of communication base stations by unmanned aerial vehicle |
| CN117808880A (en) * | 2024-01-04 | 2024-04-02 | 桂林电子科技大学 | A monocular vision spatial cooperative target pose measurement method |
Non-Patent Citations (1)
| Title |
|---|
| 张鹏: "基于海天线标定的海上单目测距方法", 电光与控制, vol. 30, no. 9, 12 September 2023 (2023-09-12), pages 36 - 41 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118967816B (en) | 2025-12-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111476841B (en) | A method and system for recognition and positioning based on point cloud and image | |
| CN110421562B (en) | Mechanical arm calibration system and calibration method based on four-eye stereoscopic vision | |
| CN112669385B (en) | Workpiece recognition and pose estimation method for industrial robots based on 3D point cloud features | |
| CN110211180A (en) | A kind of autonomous grasping means of mechanical arm based on deep learning | |
| CN113379849A (en) | Robot autonomous recognition intelligent grabbing method and system based on depth camera | |
| CN114474056B (en) | A monocular vision high-precision target positioning method for grasping operation | |
| CN112419429B (en) | Large-scale workpiece surface defect detection calibration method based on multiple viewing angles | |
| CN113269723B (en) | Unordered grabbing system for parts with three-dimensional visual positioning and manipulator cooperative work | |
| CN110136211A (en) | A workpiece positioning method and system based on active binocular vision technology | |
| CN106737665A (en) | The mechanical arm control system and implementation method matched based on binocular vision and SIFT feature | |
| Ma et al. | Binocular vision object positioning method for robots based on coarse-fine stereo matching | |
| Qiu et al. | A new formulation for hand–eye calibrations as point-set matching | |
| CN109760107B (en) | Monocular vision-based robot positioning accuracy evaluation method | |
| CN115187675B (en) | A fast positioning method for multi-shape targets based on target geometric constraints | |
| CN111583342A (en) | Target rapid positioning method and device based on binocular vision | |
| CN118744434A (en) | Automatic plugging and unplugging method of charging gun for mobile charging robot based on active visual positioning technology | |
| CN119458310A (en) | A method and system for correcting the initial visual guidance position of workpiece grasping | |
| CN114037753A (en) | Object grasping method applied to smart device, smart device and storage medium | |
| Jiang et al. | Calibration and pose measurement of a combined vision sensor system for industrial robot grasping of brackets | |
| CN112085732A (en) | Automatic marking device and method for multi-view image acquisition and observation angle of target object | |
| CN118967816B (en) | Machine vision-based methods, equipment, and media for 5G antenna pose estimation. | |
| CN117173681B (en) | A Deep Learning-Based Method and System for Identifying and Grasping Aircraft Maintenance Tools | |
| CN117367328A (en) | A spatio-temporal perception correlation filter-driven method for measuring the shape of robotic arm parts | |
| CN118351143A (en) | A low-power multi-point image displacement online tracking measurement system, method and medium | |
| CN117934592A (en) | Method, device, computer equipment and storage medium for determining placement pose parameters |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |