罐车装料口视觉跟踪系统研究  

Research on Visual Tracking System of Loading Port

作　　者：朱佩 杜芳芳 解博江 徐巧玉[2] ZHU Pei;DU Fangfang;XIE Bojiang;XU Qiaoyu(School of Intelligent Engineering,Huanghe Jiaotong University,Jiaozuo Henan 454950,China;School of Mechanical and Electrical Engineering,Henan University of Science and Technology,Luoyang Henan 471000,China)

机构地区：[1]黄河交通学院智能工程学院,河南焦作454950 [2]河南科技大学机电工程学院,河南洛阳471000

出　　处：《信息与电脑》2021年第14期72-75,共4页Information & Computer

摘　　要：针对目前罐车装料过程中存在的智能化程度低、可靠性差等问题,提出一种罐车装料口视觉跟踪系统。通过对卷积神经网络算法的研究,提出了基于SSD的罐车装料口跟踪算法。基于迁移学习对SSD神经网络进行了训练,并建立了深度学习模型,从而实现对罐车装料口的识别跟踪。实验结果表明该系统在普通环境或复杂环境下,都能够实现对进料口的稳定跟踪,具有良好的鲁棒性和实用性。Aiming at the problems of low intelligence degree and poor reliability in the current tanker loading process,a tracking system of tanker's loading port was proposed.Through the study of convolutional neural network algorithm,the tracking algorithm of tanker loading port based on SSD was proposed.In this design,SSD neural network was trained based on transfer learning,and a deep learning model was established,so as to realize the identification and tracking of tanker loading port.The experimental results show that the system can realize the stable tracking of the feed port in the ordinary environment or in the complex environment.It has good robustness and practicability.

关 键 词：罐车装料 深度学习 卷积神经网络 跟踪算法 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]