摘要
针对自动导引车(AGV)在工业自动化运输中定位速度慢,自动化效率低的问题,将全局视觉运用到车间物流的AGV定位导航中,提出一种基于深度学习的目标检测及定位方法。采用YOLO算法检测关键帧,通过ORB特征点与卡尔曼滤波结合的方式定位非关键帧,采用巴氏系数判断定位结果的准确性。相较于原有的YOLO定位算法,该算法在满足AGV定位精度要求的情况下,有效提高YOLO定位速度,提升车间自动化效率。
Automatic guided vehicle(AGV)has slow positioning speed and low automation efficiency in industrial automatic transportation.Global vision was applied to AGV positioning and navigation of workshop logistics,and a target detection and positioning method was proposed based on deep learning.Yolo algorithm was adopted to detect key frames,and non-key frame was located by ORB feature points combining with Kalman filter.The accuracy of positioning results was judged by Bhattacharyya coefficient.Compared with the original yolo positioning algorithm,this algorithm can effectively improve the Yolo positioning speed and workshop automation efficiency under the condition of meeting the requirements of AGV positioning accuracy.
作者
李斌
刘明兴
姚明杰
麻方达
李晓帆
符朝兴
LI Bin;LIU Ming-xing;YAO Ming-jie;MA Fang-da;LI Xiao-fan;FU Chao-xing(College of Mechanical and Electrical Engineering,Qingdao University,Qingdao 266071,China)
出处
《青岛大学学报（自然科学版）》
CAS
2022年第2期83-91,共9页
Journal of Qingdao University(Natural Science Edition)
关键词
定位
YOLO算法
ORB特征点
卡尔曼滤波
巴氏系数
location
Yolov algorithm
Orb feature points
Kalman filter
Bhattacharyya coefficient
