计算机科学

计算机科学 ›› 2024, Vol. 51 ›› Issue (9): 173-181.doi: 10.11896/jsjkx.230600056

• 计算机图形学&多媒体 • 上一篇    下一篇

基于YOLOv5s和双稳随机共振的夜间车辆检测算法

胡鹏飞1, 王友国1,2, 翟其清1, 颜俊2, 柏泉1   

  1. 1 南京邮电大学理学院 南京 210023
    2 南京邮电大学通信与信息工程学院 南京 210023
  • 收稿日期:2023-06-07 修回日期:2023-11-15 出版日期:2024-09-15 发布日期:2024-09-10
  • 通讯作者: 王友国(wangyg@njupt.edu.cn)
  • 作者简介:(984253044@qq.com)
  • 基金资助:
    国家自然科学基金(62071248)

Night Vehicle Detection Algorithm Based on YOLOv5s and Bistable Stochastic Resonance

HU Pengfei1, WANG Youguo1,2, ZHAI Qiqing1, YAN Jun2, BAI Quan1   

  1. 1 School of Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China
    2 School of Communication and Information Engineering,Nanjing University of Posts and Telecommunications,Nanjing 210023,China
  • Received:2023-06-07 Revised:2023-11-15 Online:2024-09-15 Published:2024-09-10
  • About author:HU Pengfei,born in 1999,postgra-duate.His main research interests include stochastic resonance and deep learning-based vehicle detection.
    WANG Youguo,born in 1968,Ph.D,professor,Ph.D supervisor.His main research interests include signal and information processing,social network information dissemination and control,stochastic resonance theory and application.
  • Supported by:
    National Natural Science Foundation of China(62071248).

PDF (PC)

501

摘要： 针对夜间车辆检测过程中光照不强导致漏检误检的问题,基于YOLOv5s和双稳随机共振提出一种改进的夜间车辆检测算法。YOLOv5s从4方面进行改进:1)在Backbone和Neck中更换细小结构,提高网络对小目标的检测能力;2)加入由坐标注意力CA和能量注意力SimAM构成的双注意力机制,提高网络对目标的特征提取能力;3)采用轻量化骨干Fasternet,减少模型参数量;4)在Head中采用WIoU损失函数,加快边界框回归损失的收敛速度。利用经典的双稳随机共振对夜间车辆数据集进行低照度图像增强,分析其有效性,并将增强后的夜间车辆图像传入改进的YOLOv5s网络进行训练。实验结果表明,相较于原始YOLOv5s,融合改进的YOLOv5s和双稳随机共振的夜间车辆检测算法在执行远景小目标以及密集遮挡的夜间车辆检测任务时具有更高的准确率和更低的漏检率。

关键词: 双稳随机共振, 低照度图像增强, YOLOv5s, 双注意力机制, 轻量化骨干

Abstract: Aiming at the problems of missed and false detection caused by weak illumination during night vehicle detection,an improved night vehicle detection algorithm is proposed based on bistable stochastic resonance and YOLOv5s.YOLOv5s is improved from four aspects,replacing small structures in Backbone and Neck to improve the detection ability of the network to small targets.A dual attention mechanism composed of coordinate attention CA and energy attention SimAM is added to improve the feature extraction ability of the network to the target.The lightweight backbone Fasternet is adopted to reduce the amount of model parameters.The WIoU loss function is used in Head to accelerate the convergence speed of bounding box regression loss.The effectiveness of the nighttime vehicle dataset is analyzed from quantitative and qualitative perspectives by using classical bistable stochastic resonance,and the enhanced nighttime vehicle images are passed into the improved YOLOv5s network for training.Experimental results show that,compared with the original YOLOv5s,the night vehicle detection algorithm combining improved YOLOv5s and bistable stochastic resonance has better accuracy and lower missed detection rate when performing long-range small targets and densely occluded night vehicle detection tasks.

Key words: Bistable stochastic resonance, Low-light image enhancement, YOLOv5s, Dual attention mechanism, Lightweight backbone

中图分类号: 

  • TP391
[1]ZHANG X Y,GAO H B,ZHAO J H,et al.Overview of deep learning intelligent driving methods[J].Journal of Tsinghua University(Science and Technology),2018,58(4):438-444.
[2]GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich FeatureHierarchies for Accurate Object Detection and Semantic Segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.New York:IEEE Press,2014:580-587.
[3]GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.New York:IEEE Press,2015:1440-1448.
[4]REN S,HE K,GIRSHICK R,et al.Faster R-CNN:towardsReal-Time Object Detection with Region Proposal Networks[J].IEEE Transactions on Pattern Analysis &Machine Intelligence,2017,39(6):1137-1149.
[5]LONG S,SONG X F,ZHANG S,et al.Vehicle detection research on improved aerial images of YOLOv5s[J].Laser Journal,2022,43(10):22-29.
[6]ZHANG H D,ZHANG R Q,TONG L,et al.Vehicle target detection method based on improved YOLOv5s[J].Journal of Chongqing University of Technology(Natural Science),2023,37(7):80-89.
[7]DENG T M,LIU X H,WANG L,et al.Vehicle detection algorithm combined with cascading attention mechanism[J].Computer Engineering and Applications,2023,59(21):141-150.
[8]ZOU Y,LONG W,LI Y Y,et al.Research on vehicle detection algorithm in low-illumination environment[J].Machine,2022,49(7):66-74.
[9]GUO K Y,WANG S D,LI X,et al.Multi-target detection of vehicles in dim scenes based on Dim env-YOLO algorithm[J].Computer Engineering,2023,49(3):312-320.
[10]BENZI R,SUTERA A,VULPIANI A.The mech-anism of stochastic resonance[J].Journal of Physics A:Mathematical and General,1981,14(11):L453-L457.
[11]LIU J,WANG Y G.Performance investigation of stochastic re-sonance in bistable systems with time-delayed feedback and three types of asymmetries[J].Physica A:Statistical Mechanics and Its Applications,2018,493:359-369.
[12]LIU J,HU B,WANG Y G.Optimum adaptive array stochastic resonance in noisy grayscale image restoration[J].Physics Letters A,2019,383(13):1457-1465.
[13]WUEHR M,BOERNER J C,PRADHAN C,et al.Stochasticresonance in the human vestibular system-Noise-induced facilitation of vestibulospinal reflexes[J].Brain Stimulation,2018,11(2):261-263.
[14]GAO W,XIAO H F.Low-light color image en-hancement based on dynamic bistable stochastic resonance[J].Chinese Journal of Liquid Crystals and Displays,2021,36(6):861-868.
[15]WEI M,HU X F,LIN M.Low-illumination image enhancement method based on tetrastable stochastic resonance[J].Chinese Journal of Liquid Crystals and Displays,2022,37(7):871-879.
[16]SUNKARA R,LUO T.No more strided convolutions or poo-ling:A new CNN building block for low-resolution images and small objects[C]//Machine Learning and Knowledge Discovery in Databases.France,2023:443-459.
[17]WANG C Y,BOCHKOVSKIY A,LIAO H Y M.YOLOv7:Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[J].arXiv:2207.02696,2022.
[18]HOU Q B,ZHOU D Q,FENG J S.Coordinate attention for effi-cient mobile network design[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.New York:IEEE Press,2021:13713-13722.
[19]YANG L X,ZHANG R Y,LI L,et al.SimAM:A Simple,Parameter-Free Attention Module for Convolutional Neural Networks[C]//Proceedings of the 38th International Conference on Machine Learning.PMLR,2021:11863-11874.
[20]LI H,LI J,WEI H,et al.Slim-neck by GSConv:A better design paradigm of detector architectures for autonomous vehicles[J].arXiv:2206.02424,2022.
[21]CHEN J,KAO S,HE H,et al.Run,Don't Walk:Chasing Hi-gher FLOPS for Faster Neural Networks[J].arXiv:2303.03667,2023.
[22]TONG Z,CHEN Y,XU Z,et al.Wise-IoU:Bounding Box Regression Loss with Dynamic Focusing Mechanism[J].arXiv:2301.10051,2023.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
渝ICP备16013121号-4
地址：重庆市渝北区洪湖西路18号    邮编：401121  
电话：023-63500828（编辑部） 023-67039612（会议/专题） 023-67039623（财务/发票）
E-mail：jsjkx12@163.com
本系统由北京玛格泰克科技发展有限公司设计开发