基于机器视觉的目标定位与机器人规划系统研究

Research on Target Positioning and Robot Planning System Based on Machine Vision

摘要:

为完成机械臂在非特定复杂背景环境下的自主抓取,通过设计RGB-D相机对场景内的物体进行实时检测,采用基于深度学习的目标检测定位方法,并对相机-机械臂-目标物体的三维标定模型进行研究。将物体的三维坐标信息通过ROS话题机制发送给机械臂,并通过moveIT编程规划抓取规划。 通过设计一套基于ROS的视觉检测和机械臂抓取系统,将计算机视觉检测技术以及机械臂运动规划抓取应用在机器人操作系统ROS平台上。实验结果表明,该系统可以实时高效地操作机器人来完成指定的控制作业,提高了系统对环境的适应能力,该系统具有抓取准确、物体识别准确率高的特点,解决了传统机械臂操控中的不足。

Abstract:

In order to complete the autonomous capture of the manipulator in a non-specific complex background environment, the object in the scene is detected in real time by designing the RGB-D camera, and the target detection and positioning method based on deep learning is adopted. The three-dimensional calibration model is studied, and the computer vision detection technology and the robot arm motion planning are applied to the robot operating system ROS platform. The experimental results show that the system can operate the robot in real time and efficiently to complete the specified control operations, and improve the system"s adaptability to the environment. The system has the characteristics of accurate grabbing and high object recognition accuracy, which solves the traditional robot arm control Insufficient.