A Robust Lightweight Network for Pedestrian Detection Based on YOLOv5-x

被引:1
|
作者
Chen, Yuanjie [1 ]
Wang, Chunyuan [1 ]
Zhang, Chi [2 ]
机构
[1] Shanghai Univ Engn Sci, Coll Elect & Elect Engn, Shanghai 201620, Peoples R China
[2] Shanghai Conservatory Mus, Dept Ethn Mus, Shanghai 200031, Peoples R China
来源
APPLIED SCIENCES-BASEL | 2023年 / 13卷 / 18期
关键词
lightweight network; YOLOv5; algorithm; weighted bidirectional feature pyramid; compact bilinear pooling;
D O I
10.3390/app131810225
中图分类号
O6 [化学];
学科分类号
0703 ;
摘要
Pedestrian detection is a crucial task in computer vision, with various applications in surveillance, autonomous driving, and robotics. However, detecting pedestrians in complex scenarios, such as rainy days, remains a challenging problem due to the degradation of image quality and the presence of occlusions. To address this issue, we propose RSTDet-Lite (a robust lightweight network) for pedestrian detection on rainy days, based on an improved version of YOLOv5-x. Specifically, in order to reduce the redundant parameters of the YOLOv5-x backbone network and enhance its feature extraction capability, we propose a novel approach named CBP-GNet, which incorporates a compact bilinear pooling algorithm. This new net serves as a new backbone network, resulting in significant parameter reduction and enhancing the fine-grained feature fusion capability of the network. Additionally, we introduce the Simple-BiFPN structure as a replacement for the original feature pyramid module based on the weighted bidirectional feature pyramid to further improve feature fusion efficiency. To enhance network performance, we integrate the CBAM attention mechanism and introduce the idea of structural reparameterization. To evaluate the performance of our method, we create a new dataset named RainDet3000, which consists of 3000 images captured in various rainy scenarios. The experimental results demonstrate that, compared with YOLOv5, our proposed model reduces the network size by 30 M while achieving a 4.56% increase in mAP. This confirms the effectiveness of RSTDet-Lite in achieving excellent performance in rainy-day pedestrian detection scenarios.
引用
收藏
页数:17
相关论文
共 50 条
  • [1] Little-YOLOv4: A Lightweight Pedestrian Detection Network Based on YOLOv4 and GhostNet
    Zheng, Hongtao
    Liu, Hong
    Qi, Wei
    Xie, Hao
    WIRELESS COMMUNICATIONS & MOBILE COMPUTING, 2022, 2022
  • [2] YOLOv5-AC: Attention Mechanism-Based Lightweight YOLOv5 for Track Pedestrian Detection
    Lv, Haohui
    Yan, Hanbing
    Liu, Keyang
    Zhou, Zhenwu
    Jing, Junjie
    SENSORS, 2022, 22 (15)
  • [3] A Pedestrian Detection Network Model Based on Improved YOLOv5
    Li, Ming-Lun
    Sun, Guo-Bing
    Yu, Jia-Xiang
    ENTROPY, 2023, 25 (02)
  • [4] Lightweight YOLOv5 Pedestrian Detection Algorithm Based on Pixel Difference Attention
    Chen, Gaoyu
    Wang, Xiaojun
    Li, Xiaohang
    Computer Engineering and Applications, 2025, 61 (01) : 291 - 299
  • [5] Lightweight Road Damage Detection Network Based on YOLOv5
    Zhao, Jingwei
    Tao, Ye
    Zhang, Zhixian
    Huang, Chao
    Cui, Wenhua
    ENGINEERING LETTERS, 2024, 32 (08) : 1708 - 1720
  • [6] A lightweight vehicles detection network model based on YOLOv5
    Dong, Xudong
    Yan, Shuai
    Duan, Chaoqun
    ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2022, 113
  • [7] Fast ship detection based on lightweight YOLOv5 network
    Zheng, Jia-Chun
    Sun, Shi-Dan
    Zhao, Shi-Jia
    IET IMAGE PROCESSING, 2022, 16 (06) : 1585 - 1593
  • [8] An Improved Lightweight Network MobileNetv3 Based YOLOv3 for Pedestrian Detection
    Zhang, Xiaxia
    Li, Ning
    Zhang, Ruixin
    2021 IEEE INTERNATIONAL CONFERENCE ON CONSUMER ELECTRONICS AND COMPUTER ENGINEERING (ICCECE), 2021, : 114 - 118
  • [9] A lightweight network face detection based on YOLOv5 Lightweight model face detection based on YOLOv5 combined with Mobilenetv2
    Xu, Bowen
    Wang, Chunmei
    Yu, Baocheng
    Xu, Wenxia
    Du, Bing
    2023 THE 6TH INTERNATIONAL CONFERENCE ON ROBOT SYSTEMS AND APPLICATIONS, ICRSA 2023, 2023, : 157 - 162
  • [10] Lightweight network for insulator fault detection based on improved YOLOv5
    Weng, Dehua
    Zhu, Zhiliang
    Yan, Zhengbing
    Wu, Moran
    Jiang, Ziang
    Ye, Nan
    CONNECTION SCIENCE, 2024, 36 (01)