微型电脑应用2025，Vol.41Issue(9)：9-12,4.
基于多传感器信息融合的袋装垃圾抓取方法研究
Research on Bagged Garbage Grasping Method Based on Multi-sensor Information Fusion
摘要
Abstract
Aiming at the difficulties of collecting and sorting household garbage in residential areas,a bagged garbage grasping method based on multi-sensor information fusion is proposed for mobile robots.Bagged garbage detection is completed by the improved YOLOv5 algorithm,and the NMS in the original network is replaced by distance intersection over Union-NMS(DI-oU-NMS)to improve the detection accuracy of the blocked target.The target size data of laser radar and time of flight(ToF)sensor are integrated to calculate the opening and closing degree of the mechanical claw,and achieve accurate grasping by taking a vertical positioning approach from directly above the target.Experimental results show that the target grasping success rate of the robot experimental platform after adopting the proposed method reached 94.5％.and the mean average precision(mAP)of the improved YOLOv5 algorithm on the self-made data set is improved by 7.3％.关键词
袋装垃圾抓取/多传感器信息融合/YOLOv5/DIoU-NMS/机器人应用Key words
bagged garbage grasping/multi-sensor information fusion/YOLOv5/DIoU-NMS/application of robot分类
信息技术与安全科学引用本文复制引用
许龙铭,曾惠霞,肖平..基于多传感器信息融合的袋装垃圾抓取方法研究[J].微型电脑应用,2025,41(9):9-12,4.基金项目
广东省普通高校重点领域专项(2022ZDZX1042) （2022ZDZX1042）
广东大学生科技创新培育专项资金资助项目(pdjh2022b0759) （pdjh2022b0759）
