一种面向工业机器人智能抓取的视觉引导技术研究

为实现工业机器人自主识别并抓取指定的目标,提出了一种基于计算机视觉引导的解决方法.该方法利用指定目标的3D数据模型,以及由两台或者多台CCD摄像机从工作场景中不同角度获取到的数字图像,经过目标姿态估算、投影计算并生成投影图像,再利用投影图像与目标真实图像进行比较,实现了目标的识别与位姿获取.实验结果表明该方法可以在较复杂的环境下识别指定目标,并获得目标的姿态信息,从而为工业机器人抓取目标提供引导.

Study on a technique of industrial robot vision guide oriented to intelligent grasp

To make the industrial robot recognize and grasp the object automatically,proposes a method based on computer vision guide. Through estimating the pose of target object,projection calculation by using the 3D data of target and its digital images captured by two or more CCD cameras from different angles in the work scene and then generating the projective image,and comparing it to its corresponding image captured from camera. This method achieves target object recognition and pose acquisition. The results sh...