现代计算机2024,Vol.30Issue(19) :7-12.DOI:10.3969/j.issn.1007-1423.2024.19.002

基于GOLD-YOLO改进YOLOv5模型道路病害检测研究

Research on road disease detection based on GOLD-YOLO improved YOLOv5 model

陈飞宇 张应迁 吴嘉懿 李睿鑫 彭良吉
现代计算机2024,Vol.30Issue(19) :7-12.DOI:10.3969/j.issn.1007-1423.2024.19.002

基于GOLD-YOLO改进YOLOv5模型道路病害检测研究

Research on road disease detection based on GOLD-YOLO improved YOLOv5 model

陈飞宇 1张应迁 2吴嘉懿 3李睿鑫 2彭良吉2
扫码查看

作者信息

  • 1. 四川轻化工大学机械工程学院,自贡 643000
  • 2. 四川轻化工大学土木工程学院,自贡 643000
  • 3. 四川轻化工大学教育与心理科学学院,自贡 643000
  • 折叠

摘要

随着机动车数量的增加和道路负荷的增大,道路病害问题日益严重,需要及时发现和识别各类病害以保障道路交通安全,但传统YOLOv5采用的FPN信息融合方式可能导致信息损失.因此,结合华为GOLD-YOLO中的Gather-and-Distribute模块对传统YOLOv5的特征融合模块进行改进.实验结果表明,优化后的YOLOv5算法在训练模型时收敛所需的迭代次数有了显著降低,从590次减少到366次,大大提高了训练速度.同时,总体的mAP@0.5也从原来的87.4%提升到了88.7%.

Abstract

With the increase in the number of motor vehicles and road loads,the problem of road diseases is becoming increas-ingly serious.It is necessary to timely detect and identify various diseases to ensure road traffic safety.However,the FPN informa-tion fusion method adopted by the traditional YOLOv5 may lead to information loss.Therefore,the traditional YOLOv5 feature fu-sion module is improved by combining the Gather and Distribution module in Huawei GOLD-YOLO.The experimental results show that the optimized YOLOv5 algorithm significantly reduces the number of iterations required for convergence during model train-ing,from 590 to 366,greatly improving training speed.Meanwhile,overall mAP@0.5 It has also increased from 87.4%to 88.7%.

关键词

YOLOv5/道路病害/目标检测/信息融合/信息损失/网络改性

Key words

YOLOv5/road diseases/target detection/information fusion/information loss/network modification

引用本文复制引用

出版年

2024
现代计算机
中大控股

现代计算机

影响因子:0.292
ISSN:1007-1423
段落导航相关论文