|
基于透视变换的机器人动态抓取技术研究
|
Abstract:
目的:针对工业上机器人抓取过程中,摄像头拍摄角度受限、容易产生角度误差等问题,提出一整套基于透视变换的空间坐标算法。方法:首先,搭建出一套类似于工业传送带的视觉定位抓取系统,介绍系统中各组成部分及其功能作用;其次,构建出斜视坐标系,利用透视变换技术完成斜视坐标系点坐标与俯视坐标系点坐标的自由变换,详细介绍计算过程及思路;最后,搭建出含有透视变换技术的机器人动态抓取系统,对上述透视变换技术与坐标的计算方法进行测试和验证。结果:实验结果表明,透视变换技术下坐标的计算符合设计预期,良好地实现了整个动态定位和抓取过程。结论:本文提出的基于透视变换的机器人动态抓取技术符合工程应用要求。
Objective: The work aims to propose a set of optimized spatial coordinate algorithms based on perspective transformation to solve the problem of limited camera shooting angle and angle errors in the process of industrial robot grasping. Method: Firstly, a set of visual positioning and grasping system which is similar to industrial conveyor belt is built to introduce each component in the system and its corresponding function. Secondly, an oblique coordinate system is constructed, and details of the transformation of point coordinates in the oblique coordinate system and view coordinate system based on perspective transformation technology are introduced. Finally, a robot dynamic grasping system with perspective transformation technology is built to test and verify the coordinate calculation method. Results: The experimental results show that the calculation of the coordinate based on the perspective transformation technology meets the design expectations, and the entire dynamic positioning and grasping process is well realized. Conclusion: It can be concluded that the robot dynamic grasping technology based on perspective transformation meets the requirements of engineering applications.
[1] | 戴雯惠, 樊凌. 基于改进透视变换的畸变图像校正方法研究[J]. 信息通信, 2020(11): 63-65. |
[2] | 彭妍, 郭君斌, 于传强, 柯冰. 基于平面变换的高精度相机标定方法[J/OL]. 北京航空航天大学学报. https://doi.org/10.13700/j.bh.1001-5965.2021.0015, 2021-04-19. |
[3] | 陈志伟, 徐世许, 刘云鹏, 曾祥晓. 基于视觉筛选的并联机器人平面抓取系统设计[J]. 制造业自动化, 2018, 40(5): 44-47. |
[4] | 黄金梭, 沈正华. 基于视觉和工业机器人的动态抓取技术[J]. 包装工程, 2019, 40(11): 177-182. |
[5] | 张云龙, 王清珍. 基于视觉的多自主移动机器人动态编队技术研究[J]. 人工智能与机器人研究, 2021,10(1): 29-36. |
[6] | 邱素贞, 李庆年, 卢志翔, 阮承海. 基于机器视觉检测的码垛机器人控制系统设计[J]. 包装工程, 2019, 40(3): 207-211. |