基于双目视觉的人形机器人物体定位与抓取

Object Location and Grasping of Humanoid Robot Based on Binocular Vision

以人形机器人在非结构化的环境中识别、目标定位与实物抓取为基础,使用仿人机器人NAO,进行双目视觉模块扩展,对机器人手臂进行运动学建模、双目视觉标定、视觉识别与定位、视觉伺服控制等问题分析,并通过一系列实验验证了NAO机器人可以准确有效地完成目标物体识别、定位与抓取任务.对服务型机器人的应用与深度开发具有参考意义.

Based on humanoid robot′s recognition,target positioning and object grasping in unstructured environment,the binocular vision module is expanded by using the humanoid robot Nao.The robot arm kinematics modeling,binocular vision calibration,visual recognition and positioning,visual servo control and other issues are analyzed,and the NAO is verified by a series of experiments.The robot can accurately and effectively complete the tasks of target recognition,positioning and grasping.It is of great reference to the application and deep development of service robots.