摘要
为了克服人工遥操作的缺点,需要提高遥控焊接机器人的智能水平。准确的任务空间三维建模是遥控机器人规划、运动控制和其它自动化功能的基础。使用立体视觉技术对遥控焊接进行任务空间分析和三维重建。其中的立体匹配利用人机交互并结合亚像素角点特征提取算法完成。该算法先检测Harris角点,然后进行亚像素优化,克服了视觉自动建模中计算复杂、鲁棒性差的缺点。任务空间三维重建的试验结果表明了该技术的可行性。
In order to overcome the disadvantage of the human tele-operation, there is need to improve the intelligence level of the tele-robotic welding system. Accurate 3D modeling of the task space is the essential foundation for telerobotic planning, motion control and other automatic function. Stereo vision is employed for task space analysis and 3D reconstruction of telerobotic welding. Stereo matching is accomplished by man-machine interaction combined with subpixel corner feature detection algorithm, in which the Harris corner is first detected then the subpixel refine is made. So that the complicated calculation and low robustness of the automatic visual modeling algorithm is avoided. The experimental 3D reconstruction results of task space have shown the validity of the technique.
出处
《焊接学报》
EI
CAS
CSCD
北大核心
2007年第11期5-8,共4页
Transactions of The China Welding Institution
基金
总装"十一五"预研项目
哈尔滨工业大学优秀团队支持计划资助项目
