智能系统学报2025，Vol.20Issue(2)：435-444,10.DOI:10.11992/tis.202403021
图像信息输入下的软线重构与抓取验证
Flexible wire reconstructing and gripping verification under the image information input
摘要
Abstract
To address the complex task of operating flexible objects with a robot,a concise and efficient method for grasping soft wires is designed.First,the soft wire is measured using a stereo camera and image processing techniques.By combining the transformation relationships between coordinate systems,three-dimensional coordinate information is reconstructed from the image data.The soft wire is then modeled and reconstructed using the Frenet frame,enabling the acquisition of positional and bending direction information at any arbitrary point along the wire.The representations and transformation methods for the six-DOF robotic arm pose,along with the camera calibration process,are provided.These enable the conversion of the Frenet frame's coordinates and directional data for any point along the curve into the robotic gripper's pose.Finally,a ground experiment was conducted to demonstrate the robot's ability to grasp the soft wire along the bending direction of a target point,verifying the feasibility of this method.关键词
软线/视觉感知/深度相机/图像处理/坐标计算/曲线重构/位姿转换/机械臂抓取Key words
flexible wire/visual sensing/depth camera/image processing/coordinates calculation/curve reconstruction/pose transformation/robot arm grasping分类
计算机与自动化引用本文复制引用
邱介禄,王明明..图像信息输入下的软线重构与抓取验证[J].智能系统学报,2025,20(2):435-444,10.基金项目
国家自然科学基金项目(61973256) （61973256）
广东省基础与应用基础研究基金项目(2024A1515011178). （2024A1515011178）
