| 引用本文: | 黄梦洁,叶磊,易凡骁,王千.面向机器人控制的直接视觉伺服技术发展综述[J].控制理论与应用,2025,42(9):1681~1699.[点击复制] |
| HUANG Meng-jie,YE Lei,YI Fan-xiao,WANG Qian.A review of the development of direct vision servo technology for robot control[J].Control Theory & Applications,2025,42(9):1681~1699.[点击复制] |
|
| 面向机器人控制的直接视觉伺服技术发展综述 |
| A review of the development of direct vision servo technology for robot control |
| 摘要点击 4841 全文点击 324 投稿时间：2023-04-20 修订日期：2025-03-08 |
| 查看全文 查看/发表评论 下载PDF阅读器 HTML |
| DOI编号 10.7641/CTA.2024.30248 |
| 2025,42(9):1681-1699 |
| 中文关键词 机器人控制 视觉伺服 直接视觉伺服 图像特征描述方法 相似性度量 优化控制策略 |
| 英文关键词 robot control visual servo direct visual servo description methods of image features similarity measures optimize control strategies |
| 基金项目 |
|
| 中文摘要 |
| 视觉伺服作为一种利用图像信息来实现机器人闭环控制的方法,在移动机器人导航控制和机械臂高精度
控制等方面有广泛的应用.相较于基于局部特征的经典视觉伺服方法,直接视觉伺服技术利用了更多的图像全局
信息,在控制系统的鲁棒性和收敛性上有较大优势,因此本文重点从视觉处理和控制器设计两部分介绍直接视觉伺
服技术.本文首先介绍了经典视觉伺服框架,然后研究了直接视觉伺服技术中的视觉特征描述方法、特征相似性度
量方法,并对比了不同特征描述方法在收敛域和运算量等性能方面的差异,之后研究了直接视觉伺服中基于非线性
优化等优化策略的控制器设计方法,最后展望了深度学习在直接视觉伺服中的作用. |
| 英文摘要 |
| Visual servo is a method that uses image information to realize closed-loop control of robots, which has
a wide range of applications in mobile robot navigation control and high-precision control of robotic arms. Compared
with the classical visual servo method based on local features, direct visual servo technology makes more use of image
global information, which has great advantages in the robustness and convergence of the control system. So this paper
focuses on direct visual servo technology from two parts: visual processing and controller designing. This paper first
introduces the classical visual servo framework, then studies the visual feature description method and the feature similarity
measurement method in direct visual servo technology, and compares the differences in the performance of different feature
description methods in convergence domain and computing amount, and then studies the controller design methods based
on optimization strategy like nonlinear optimization in direct visual servo, and finally looks forward to the role of deep
learning in direct visual servo. |
|
|
|
|
|