期刊文献+

基于改进YOLOv4的道路标线检测算法

Road Marking Detection Algorithm Based on Improved YOLOv4
下载PDF
导出
摘要 道路标线在维护交通秩序、保证交通通畅方面具有重要作用,所以保证道路标线的完整性具有重要意义。本文提出了一种基于改进YOLOv4目标检测算法的道路标线检方法。本文使用深度可分离网络降低YOLOv4的参数量,同时使用了Triplet注意力机制增强了网络的表达能力,为加速整个模型收敛速度同时使用EIOU优化损失函数,极大的增加了模型的收敛速度。结果表明更改后的模型的性能更佳,其预测精度达到了97%,而参数量为只有11.703M,且只有原来的1/4。 Road markings play an important role in maintaining traffic order and ensuring smooth traffic, so ensuring the integrity of road markings is of great significance. Therefore, this paper proposes a road marking detection method based on improved YOLOv4 target detection algorithm. In this paper, the depth separable network is used to reduce the number of parameters of YOLOv4, and the Triplet attention mechanism is used to enhance the expression ability of the network. In order to accelerate the convergence speed of the whole model, EIOU is used to optimize the loss function, which greatly increases the convergence speed of the model. The results show that the performance of the modified model is better, its prediction accuracy reaches 97%, and the parameter quantity is only 11.703M, and only 1/4of the original.
作者 赵明升 来跃深 ZHAO Ming-sheng;LAI Yue-shen(School of Mechanical and Electrical Engineering,Xi'an University of Technology,Xi'an 710021,China)
出处 《价值工程》 2022年第30期157-159,共3页 Value Engineering
基金 陕西省重点研发计划项目(2020GY-160)。
关键词 深度学习 注意力机制 轻量化 YOLOv4 EIOU deep learning attention mechanism lightweight YOLOv4 EIOU
  • 相关文献

参考文献1

二级参考文献5

共引文献11

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部