论文部分内容阅读
以ABB机器人、快换手爪、双目相机为硬件基础,搭建了基于双目立体视觉的工业机器人智能抓取系统,旨在实现随机放置物体的识别定位和自动抓取。基于视觉模块,研究了图像处理、立体匹配和深度计算以及坐标转换,以C++为开发平台,结合Open CV图像开发库和Triclops库,开发了视觉识别与定位算法,实现对圆柱体和长方体薄板对象的分类和位姿识别。最后,通过实验验证,表明该视觉算法有较高的识别定位精度,满足抓取系统的要求。
Based on the hardware of ABB robots, quick-change claws and binocular cameras, an intelligent robotic grab system based on binocular stereo vision was set up to realize the recognition and positioning of random objects and the automatic grasping. Based on the visual module, image processing, stereo matching, depth calculation and coordinate transformation are studied. With C ++ as the development platform, a visual recognition and localization algorithm is developed based on Open CV image library and Triclops library. Classification and pose recognition. Finally, the experimental verification shows that the visual algorithm has high recognition accuracy and meets the requirements of the crawler system.