基于双目视觉的机器人目标定位与机械臂控制

摘要

In order to interact with complex and changeable environments,finish the recognition and grab for object,a humanoid robot arm control method based on binocular vision is presented that employed in service robot.Manipulator is modeled by D-H method in this paper,the model is improved,a sample inverse arithmetic of humanoid manipulator is presented.The target object is recognized by color segmentation and the 3D position is computed by the stereo vision system.After positioning the target,a blind grasp is performed by the manipulator.Experimental results are conducted for illustration of effectiveness of the proposed methods,and the service robot can recognize and grab the object.%为了更好地与复杂多变的非结构化环境进行交互,完成对目标物体的识别和抓取,提出了一种应用于服务机器人平台的基于双目视觉的仿人机械臂控制方法.文中首先用D-H方法对机械臂进行建模,并对这个模型做了改进,给出了一种更加简便的3+1自由度仿人机械臂的逆解算法,采用基于双目视觉与颜色分割的目标识别方法;然后根据识别出的目标三维坐标信息控制机械臂完成抓取任务;最后,本方法在家庭服务机器人上得到了验证,机器人能够完成对目标物体的识别和抓取动作.