基于双目视觉的食品分拣Delta机器人定位抓取技术  

Food sorting Delta robot positioning and grasping technology based on binocular vision

作　　者：米西峰[1] 尚展垒 范媛媛[1] 孟晓川 MI Xi-feng;SHANG Zhan-lei;FAN Yuan-yuan;MENG Xiao-chuan(Jiaozuo Normal College,Jiaozuo,Henan 454000,China;Zhengzhou University of Light Industry,Zhengzhou,Henan 450002,China;Yanshan University,Qinhuangdao,Hebei 066004,China)

机构地区：[1]焦作师范高等专科学校,河南焦作454000 [2]郑州轻工业大学,河南郑州450002 [3]燕山大学,河北秦皇岛066004

出　　处：《食品与机械》2022年第6期112-116,212,共6页Food and Machinery

基　　金：河南省软科学研究项目(编号:212400410223)。

摘　　要：目的:解决Delta机器人在食品分拣中定位精度差、抓取成功率低等问题,提高机器人的柔性抓取能力。方法:在现有双目视觉和Delta机器人技术的基础上,提出一种基于双目视觉的Delta机器人柔性抓取方法。采用张正友标定方法对相机进行标定,通过弦中点Hough变换对目标进行识别和定位,根据PID跟踪抓取控制Delta机器人抓取动态目标。通过试验对单目标分拣和多目标分拣性能进行分析,验证该方法的可行性。结果:该方法能够准确、快速、稳定地对目标进行动态抓取(成功率在94.0%以上)。结论:双目视觉与Delta机器人相结合可以有效提高机器人的定位精度和分拣抓取的成功率。Objective:Solve the problems of poor positioning accuracy and low grasping success rate of the Delta robot in food sorting,and improve the robot’s flexible grasping ability.Methods:Based on the existing binocular vision and delta robot technology,a flexible grasping method of delta robot based on binocular vision was proposed.Zhang Zhengyou calibration method was used to calibrate the camera,the target was identified and located by Hough transform at the chord midpoint,and the delta robot was controlled to grasp the dynamic target according to PID tracking and grasping.The feasibility of this method was verified by analyzing the performance of single target sorting and multi-target sorting.Results:This method can accurately,quickly and stably grab the target dynamically(success rate was above 94.0%).Conclusion:The combination of binocular vision and delta robot can effectively improve the positioning accuracy of the robot and the success rate of sorting and grasping.

关 键 词：Delta机器人 食品分拣 双目视觉 HOUGH变换 PID跟踪抓取 

分 类 号：TS205[轻工技术与工程—食品科学] TP242[轻工技术与工程—食品科学与工程] TP391.41[自动化与计算机技术—检测技术与自动化装置]