系统仿真学报 ›› 2016, Vol. 28 ›› Issue (6): 1255-1261.

• 仿真建模理论与方法 • 上一篇    下一篇

基于2D重构特征的模型化视觉伺服方法研究

张国亮, 杜吉祥, 王展妮, 王田   

  1. 华侨大学计算机科学与技术学院,厦门 361021
  • 收稿日期:2015-07-08 修回日期:2015-10-15 出版日期:2016-06-08 发布日期:2020-06-08
  • 作者简介:张国亮(1978-),男,辽宁,博士,讲师,研究方向为机器人视觉伺服、机器人遥操作。
  • 基金资助:
    国家自然科学基金(61175121,61202299),福建省自然科学基金(2016J01302,2013J06014); 华侨大学中青年教师科研提升资助计划(ZQN-YX108)

Study on Model-based Visual Servoing with Reconstructed 2D Features

Zhang Guoliang, Du Jixiang, Wang Zhanni, Wang Tian   

  1. Computer Science and Technology Institute, Huaqiao University, Xiamen 361021, China
  • Received:2015-07-08 Revised:2015-10-15 Online:2016-06-08 Published:2020-06-08

摘要: 针对传统基于图像的视觉伺服在处理摄像机退化等方面存在的不足,研究了融合三维特征与二维特征的图像视觉伺服方法。通过在经典IBVS(Image based visual servoing)控制结构基础上,将二维图像特征重构为三维特征,建立了模型化的3D特征机器人视觉伺服控制模型,并通过Simulink等仿真工具,分析比较了2种方法在图像空间和笛卡尔空间的空间运动特性,试验结果证实了方法的可行性和有效性。

关键词: 机器人视觉伺服, 摄像机退化, IBVS, 特征重构

Abstract: To overcome the defect of camera retreat in image based visual servoing, a model-based visual servoing incorporated 2D image based visual servoing with 3D features was proposed. The robot visual servoing model was established by altering traditional two dimensional feature to three dimensional features based on traditional IBVS control structure. Furthermore, combined with the Simulink software, simulation model was established and motion characteristics of two methods were analyzed and compared in image space and Cartesian space. The experiment results confirm the feasibility and effectiveness of the proposed method.

Key words: robot visual servoing, camera retreat, image based visual servoing, feature reconstruction

中图分类号: