辽宁石油化工大学学报

辽宁石油化工大学学报 ›› 2022, Vol. 42 ›› Issue (6): 90-96.DOI: 10.3969/j.issn.1672-6952.2022.06.015

• 信息与控制工程 • 上一篇    

基于孪生网络的协作机器人目标追踪

韩江雪1(), 郭小明2(), 汤永恒1, 王丽鑫2, 潘斌2   

  1. 1.辽宁石油化工大学 计算机与通信工程学院,辽宁 抚顺 113001
    2.辽宁石油化工大学 理学院, 辽宁 抚顺 113001
  • 收稿日期:2021-04-12 修回日期:2022-12-04 出版日期:2022-12-25 发布日期:2023-01-07
  • 通讯作者: 郭小明
  • 作者简介:韩江雪(1996⁃),男,硕士研究生,从事计算机视觉、图像处理方面的研究;E⁃mail:18641365626@163.com
  • 基金资助:
    国家自然科学基金项目(61602228);辽宁省“兴辽英才计划”青年拔尖人才项目(XLYC1807266);辽宁省自然科学基金项目(2015020041)

Cooperative Robot Object Tracking Based on Siamese Network

Jiangxue Han1(), Xiaoming Guo2(), Yongheng Tang1, Lixin Wang2, Bin Pan2   

  1. 1.School of Computer and Communication Engineering,Liaoning Petrochemical University,Fushun Liaoning 113001,China
    2.College of Sciences,Liaoning Petrochemical University,Fushun Liaoning 113001,China
  • Received:2021-04-12 Revised:2022-12-04 Published:2022-12-25 Online:2023-01-07
  • Contact: Xiaoming Guo

摘要:

利用自身高速高精度的特点,协作机器人通过模仿人的创造性复杂动作来提高生产效率。当前协作机器人对人动作的模仿主要来自部署人员的长期调试,缺少通用的解决方案,无法快速部署。基于此,提出了一种无锚的基于RepVGG网络的孪生网络协作机器人目标跟踪算法。该算法由孪生网络模块、分类回归模块和机器人执行模块组成。孪生网络模块使用改进的RepVGG网络代替主流的ResNet作为骨干网络用于图片特征的提取,在不损失精度的前提下提高整个网络的运行速度,降低算法对硬件的要求,对专用深度学习芯片更加友好;分类回归模块通过引入中心度分支来提高跟踪框的中心点预测精度;机器人执行模块采用尺度惩罚和宽高比惩罚以平滑跟踪框,保证协作机器人的动作流畅。实验结果表明,平均速率相比替代ResNet骨干网络前提高了14 FPS,实现了实时跟踪的效果。

关键词: 目标跟踪, 孪生网络, RepVGG网络, 深度学习, 协作机器人

Abstract:

Taking advantage of its high?speed and high?precision characteristics, cooperative robots can improve production efficiency by imitating human creative and complex actions. At present, the simulation of human action mainly comes from the long?term debugging of the deployment personnel, which is lack of general solutions and can′t be deployed quickly. Based on this, an anchor?free RepVGG network?based Siamese network collaborative robot target tracking algorithm was proposed. The algorithm consists of a siamese network module, a classification regression module and a robot execution module. The siamese network module used the improved RepVGG network instead of ResNet as the backbone network to extract image features, which can improve the running speed of the whole network without losing accuracy,reduce the hardware requirements and is more friendly to special deep learning chips; the classification and regression module introduced the centrality branch to improve the prediction accuracy of the center point of the tracking frame; the robot execution module uses scale penalty and aspect ratio penalty to smooth the tracking boxes and ensure the smooth operation of the collaborative robots. Experimental results show that the average rate is 14 FPS higher than that before replacing the backbone network, which realizes real?time object tracking.

Key words: Object tracking, Siamese network, RepVGG network, Deep learning, Cooperative robot

中图分类号: 

引用本文

韩江雪, 郭小明, 汤永恒, 王丽鑫, 潘斌. 基于孪生网络的协作机器人目标追踪[J]. 辽宁石油化工大学学报, 2022, 42(6): 90-96.

Jiangxue Han, Xiaoming Guo, Yongheng Tang, Lixin Wang, Bin Pan. Cooperative Robot Object Tracking Based on Siamese Network[J]. Journal of Liaoning Petrochemical University, 2022, 42(6): 90-96.

使用本文

0
    /   /   推荐

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://journal.lnpu.edu.cn/CN/10.3969/j.issn.1672-6952.2022.06.015

               http://journal.lnpu.edu.cn/CN/Y2022/V42/I6/90