Object position/pose estimation using CAD models for navigation of manipulator with a single CCD camera

Object position/pose estimation using CAD models for navigation of manipulator with a single CCD camera
复制标题

使用 CAD 模型进行物体位置/姿态估计,通过单个 CCD 相机对机械臂进行导航

DOI:
10.1109/cira.2003.1222208
复制
发表时间:
2003
期刊:
Proceedings 2003 IEEE International Symposium on Computational Intelligence in Robotics and Automation. Computational Intelligence in Robotics and Automation for the New Millennium (Cat. No.03EX694)
影响因子:
--
通讯作者:
Toshiya Hayashi
Toshiya Hayashi
中科院分区:
--
文献类型:
--
作者:
K. Nakao;K. Kondo;Syoji Kobashi;Y. Hata;T. Yagi;Toshiya Hayashi

文献摘要

被引文献

相似文献

提出了一种利用单摄像机和三维CAD模型进行位姿识别和机械手导航到期望位置的方法。在位姿估计中,不是同时估计目标物体的所有位姿参数,而是根据图像特征分别估计这些参数。因此,我们的方法包括位置参数估计、位姿参数估计和机械手导航到期望位置。首先利用图像特征近似估计位置参数,然后估计姿态参数。然后,基于估计将机械手移动到所需位置。重复估计和运动,直到机械手到达所需位置。这种渐进的方法实现了机械手的导航,涉及自动化工作,速度和精度都有所提高。
We propose the method for position/pose identification and manipulator navigation to the desired position using one camera and three-dimensional CAD model. In position/pose estimation, instead of simultaneous estimation of all position/pose parameters of target object, the parameters are severally estimated with image features. Thus, our method consists of position parameters estimation; pose parameters estimation and manipulator navigation to the desired position. The position parameters are estimated approximately with image features, and then pose parameters are estimated. And then, a manipulator to the desired position based on the estimates is moved. Both estimation and movement are repeated until the manipulator reaches to the desired position. This gradual approach achieves manipulator navigation concerning automation works with improvement in speed and accuracy.