基于Kinova机械臂RGBD相机的目标抓取  

Target Grasping Based on Kinova Robot Arm and RGBD Camera

作　　者：黄伟 郭瑞科[2] 刘大勇 康柯 周忠达 胡海东[2] HUANG Wei;GUO Ruike;LIU Dayong;KANG Ke;ZHOU Zhongda;HU Haidong(China Academy of Space Technology Hangzhou Institute,Hangzhou 310024,China;Beijing Institute of Control Engineering,Beijing 100094,China)

机构地区：[1]中国空间技术研究院杭州中心,杭州310024 [2]北京控制工程研究所,北京100094

出　　处：《空间控制技术与应用》2024年第1期100-106,共7页Aerospace Control and Application

基　　金：浙江省“尖兵”“领雁”计划项目(2023C01177)。

摘　　要：在轨服务与维护是未来航天技术的重要发展方向.在轨服务如燃料加注、剪切与操作电线、旋拧盖子和拆装阀门等任务,具有负载轻、近距离感知及精准定位和抓取等特点.针对这些应用场景,采用轻量化机械臂Kinova及自带彩色和深度(red,green,blue and depth,RGBD)相机进行感知、定位和抓取,完成地面验证实验,为在轨服务提供理论支持.介绍了机械臂抓取系统的研究背景并描述开发平台及其特点,以及相机标定、彩色相机与深度相机配准原理.利用机器人操作系统(robot operating system,ROS)的坐标变换树得到彩色图像目标像素在世界坐标系的坐标位置.针对特定的抓取对象提出基于色调、饱和度和亮度(hue,saturation and value,HSV)颜色空间图像分割方法,进行目标检测、分割以及空间定位,实验环节机械臂根据视觉定位结果完成目标抓取过程,证明了颜色空间图像分割方法和深度图结合定位的有效性,为在轨服务提供了理论验证.On orbit service and maintenance is an important development direction of space technology in the future.In orbit services such as fuel refueling,cutting and operating wires,screwing caps,and disassembling valves have the characteristics of light load,close range sensing,precise positioning,and grasping.For these application scenarios,this article uses lightweight robotic arm Kinova and built-in RGBD(red,green,blue,depth)cameras for perception,positioning,grasping,and completing ground verification experiments to provide theoretical support for in orbit services.This article first introduces the research background of the robotic arm grasping system and describes the development platform and its characteristics.Secondly,introduce camera calibration,color camera and depth camera registration principles,and use the TF(transform)tree of ROS(robot operating system)to obtain three-dimensional coordinates.A color space image segmentation method based on HSV(Hue,Saturation,Value)is proposed for specific captured objects,for object detection,segmentation,and spatial positioning.Finally,the experimental stage involves the robotic arm completing the target capture process based on the visual positioning results.The effectiveness of combining color space image segmentation methods with depth maps for localization has been demonstrated,providing theoretical validation for in orbit services.

关 键 词：机器臂 相机标定 RGBD相机 HSV 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]