广西师范大学学报（自然科学版） ›› 2025, Vol. 43 ›› Issue (3): 72-83.doi: 10.16088/j.issn.1001-6600.2024120101

• 智能信息处理 • 上一篇    下一篇

基于YOLOv8的雾天车辆行人实时检测方法

汤亮*, 陈博文, 牛一森, 马荣庚   

  1. 湖北工业大学机械工程学院,湖北武汉 430068
  • 收稿日期:2024-12-01 修回日期:2024-12-23 出版日期:2025-05-05 发布日期:2025-05-14
  • 通讯作者: 汤亮(1978—),男,湖北十堰人,湖北工业大学副教授,博士。E-mail:tangliang@hbut.edu.cn
  • 基金资助:
    国家科技支撑计划项目(2016IM020200-01)

A YOLOv8-based Real-time Object Detection Method for Vehicles and Pedestrians in Foggy Weather

TANG Liang*, CHEN Bowen, NIU Yisen, MA Ronggeng   

  1. School of Mechanical Engineering, Hubei University of Technology, Wuhan Hubei 430068, China
  • Received:2024-12-01 Revised:2024-12-23 Online:2025-05-05 Published:2025-05-14

PDF (PC)

172

摘要： 随着智能通信技术在智能交通场景的广泛运用,行人、车辆目标检测已成为保障道路安全的重要基础。针对在雾天恶劣环境中检测网络漏检率高、检测速度慢的问题,本文提出基于YOLOv8的实时雾天目标检测方法。该模型将输入图片加入去雾网络模块对输入图像进行预处理,保留原图片的细节特征并去除雾气的遮挡,再使用改进后的YOLOv8n进行检测。在YOLOv8n上基于FasterNet改进C2f模块,降低模型参数量及模型大小,增加模型计算效率,并设计SE-ResNeXt检测头,避免了因堆积神经网络层数带来的负面影响。最后运用知识蒸馏的方式,进一步提高检测精度。将所提出模型在 reside rtts数据集和合成有雾数据集上进行验证。与原网络相比,平均精度(mAP@50_95)提升5.2个百分点,检测帧数达到170 frame/s。

关键词: 雾天场景, 目标检测, 信息交互, FasterNet, SENet, ResNeXt

Abstract: With the extensive application of intelligent communication technology in smart traffic scenarios, the task of detecting pedestrians and vehicles constitutes an important technical means for road safety. In light of the high missed detection rate and slow detection speed in the foggy environment, a real-time foggy target detection method based on YOLOv8 is proposed. The model incorporates the fog removal network module into the input image to preprocess it, retains the detailed features of the original image and eliminates the obstruction of fog, and then utilizes the improved YOLOv8n for detection. On YOLOv8n, the C2f module is enhanced based on FasterNet to reduce the model parameters and size, increase the model’s computing efficiency, and the SE-ResNeXt detection head is designed to avoid the negative impacts of stacking neural network layers. Finally, knowledge distillation is employed to further enhance the detection accuracy. The proposed model is validated on the RTTS dataset and the synthetic foggy dataset. Compared with the original network, the average precision (mAP@50_95) is improved by 5.2 percentage points, and the detection frame rate reaches 170 frame/s.

Key words: foggy scene, object detection, information interactive;Fasternet, SE-Net, ResNeXt

中图分类号:  TP391.41

[1] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2016: 779-788. DOI: 10.1109/CVPR.2016.91.
[2]LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector[C]// Computer Vision-ECCV 2016. Cham: Springer, 2016: 21-37. DOI: 10.1007/978-3-319-46448-0_2.
[3]LI B, CHEN Y H, XU H, et al. Fast vehicle detection algorithm based on lightweight YOLO7-tiny[EB/OL]. (2023-04-17)[2024-12-01]. https://arxiv.org/abs/2304.06002. DOI: 10.48550/arXiv.2304.06002.
[4]GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]// 2014 IEEE Conference on Computer Vision and Pattern Recognition. Los Alamitos, CA: IEEE Computer Society, 2014: 580-587. DOI: 10.1109/CVPR.2014.81.
[5]REN S Q, HE K M, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149. DOI: 10.1109/TPAMI.2016.2577031.
[6]WANG J W, CHEN C, WANG C W. Street sign recognition algorithm based on deep learning[C]// Proceedings of the 2020 3rd International Conference on Image and Graphics Processing. New York: Association for Computing Machinery, 2020: 31-35. DOI: 10.1145/3383812.3383819.
[7]李轩,李静,王海燕.复杂气象条件下的交通场景目标检测算法研究[J].计算机仿真,2021,38(2):87-90,105. DOI: 10.3969/j.issn.1006-9348.2021.02.020.
[8]LI G F, YANG Y F, QU X D. Deep learning approaches on pedestrian detection in hazy weather[J]. IEEE Transactions on Industrial Electronics, 2020, 67(10): 8889-8899. DOI: 10.1109/TIE.2019.2945295.
[9]LI B Y, PENG X L, WANG Z Y, et al. AOD-Net: all-in-one dehazing network[C]// 2017 IEEE International Conference on Computer Vision (ICCV). Los Alamitos, CA: IEEE Computer Society, 2017: 4780-4788. DOI: 10.1109/ICCV.2017.511.
[10]院老虎,翟柯嘉,张泽鹏,等.基于模拟雾天遥感数据集的飞机目标检测研究[J].南京邮电大学学报(自然科学版),2021,41(3):77-84. DOI: 10.14132/j.cnki.1673-5439.2021.03.011.
[11]MA Y H, CAI J F, TAO J X, et al. Foggy image detection based on DehazeNet with improved SSD[C]// Proceedings of the 2021 5th International Conference on Innovation in Artificial Intelligence. New York: Association for Computing Machinery, 2021: 82-86. DOI: 10.1145/3461353.3461363.
[12]HAN X F. Modified cascade RCNN based on contextual information for vehicle detection[J]. Sensing and Imaging, 2021, 22(1): 19. DOI: 10.1007/s11220-021-00342-6.
[13]董丽丽,丁畅,许文海.基于直方图均衡化图像增强的两种改进方法[J].电子学报,2018,46(10):2367-2375. DOI: 10.3969/j.issn.0372-2112.2018.10.009.
[14]RAIKWAR S C, TAPASWI S, CHAKRABORTY S. Bounding function for fast computation of transmission in single image dehazing[J]. Multimedia Tools and Applications, 2022, 81(4): 5349-5372. DOI: 10.1007/s11042-021-11752-9.
[15]JU M Y, DING C, REN W Q, et al. IDE: image dehazing and exposure using an enhanced atmospheric scattering model[J]. IEEE Transactions on Image Processing, 2021, 30: 2180-2192. DOI: 10.1109/TIP.2021.3050643.
[16]HE K M, SUN J, TANG X O. Single image haze removal using dark channel prior[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2011, 33(12): 2341-2353. DOI: 10.1109/TPAMI.2010.168.
[17]LI S Y, LIN J, YANG X, et al. BPFD-Net: enhanced dehazing model based on pix2pix framework for single image[J]. Machine Vision and Applications, 2021, 32(6): 124. DOI: 10.1007/s00138-021-01248-9.
[18]陈清江,张雪.混合残差学习与导向滤波算法在图像去雾中的应用[J].光学精密工程,2019,27(12):2702-2712. DOI: 10.3788/OPE.20192712.2702.
[19]CHEN J R, KAO S H, HE H, et al. Run, don’t walk: chasing higher FLOPS for faster neural networks[C]// 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2023: 12021-12031. DOI: 10.1109/CVPR52729.2023.01157.
[20]LIU G L, SHIH K J, WANG T C, et al. Partial convolution based padding[EB/OL]. (2018-11-28)[2024-12-01]. https://arxiv.org/abs/1811.11718. DOI: 10.48550/arXiv.1811.11718.
[21]孙锦,尹明锋,谢涛,等.基于多尺度特征融合的雾天目标检测方法[J].激光杂志,2024,45(6):106-113. DOI: 10.14016/j.cnki.jgzz.2024.06.106.
[22]IOFFE S, SZEGEDY C. Batch normalization: accelerating deep network training by reducing internal covariate shift[EB/OL]. (2015-03-02)[2024-12-01]. https://arxiv.org/abs/1502.03167. DOI: 10.48550/arXiv.1502.03167.
[23]CHEN Y H, TANG X, QI X B, et al. Learning graph normalization for graph neural networks[J]. Neurocomputing, 2022, 493: 613-625. DOI: 10.1016/j.neucom.2022.01.003.
[24]娄铮铮,张欣,胡世哲,等.一种基于YOLOX_s的雾天场景目标检测方法[J].计算机科学,2024,51(7):206-213. DOI: 10.11896/jsjkx.230400086.
[25]HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2016: 770-778. DOI: 10.1109/CVPR.2016.90.
[26]XIE S N, GIRSHICK R, DOLLÁR P, et al. Aggregated residual transformations for deep neural networks[C]// 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2017: 5987-5995. DOI: 10.1109/CVPR.2017.634.
[27]SZEGEDYC, IOFFE S, VANHOUCKE V, et al. Inception-v4, inception-ResNet and the impact of residual connections on learning[EB/OL]. (2016-08-23)[2024-12-01]. https://arxiv.org/abs/1602.07261. DOI: 10.48550/arXiv.1602.07261.
[28]HU J, SHEN L, ALBANIE S, et al. Squeeze-and-excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8): 2011-2023. DOI: 10.1109/TPAMI.2019.2913372.
[29]YANG Z D, LI Z, SHAO M Q, et al. Masked generative distillation[C]// Computer Vision-ECCV 2022. Cham: Springer, 2022: 53-69. DOI: 10.1007/978-3-031-20083-0_4.
[30]院老虎,常玉坤,刘家夫.基于改进YOLOv5s的雾天场景车辆检测方法[J].郑州大学学报(工学版),2023,44(3): 35-41. DOI: 10.13705/j.issn.1671-6833.2023.03.005.
[1] 黄仁慧, 张锐锋, 文晓浩, 闭金杰, 黄守麟, 李廷会. 基于复数协方差卷积神经网络的运动想象脑电信号解码方法[J]. 广西师范大学学报（自然科学版）, 2025, 43(3): 43-56.
[2] 梁胤杰, 南新元, 蔡鑫, 李云鹏, 勾海光. 基于数据增广与改进YOLOv8的桥梁缺陷检测[J]. 广西师范大学学报（自然科学版）, 2025, 43(3): 84-97.
[3] 郭翔羽, 石天怡, 陈燕楠, 南新元, 蔡鑫. 基于YOLO-CDBW模型的列车接触网异物检测研究[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 56-69.
[4] 刘玉娜, 马双宝. 基于改进YOLOv8n的轻量化织物疵点检测算法[J]. 广西师范大学学报（自然科学版）, 2025, 43(2): 83-94.
[5] 涂智荣, 凌海英, 李帼, 陆声链, 钱婷婷, 陈明. 基于改进YOLOv7-Tiny的轻量化百香果检测方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(5): 79-90.
[6] 吕辉, 吕卫峰. 基于改进YOLOv5的眼底出血点检测算法[J]. 广西师范大学学报（自然科学版）, 2024, 42(3): 99-107.
[7] 曾亮, 胡谦, 杨腾飞, 谭微微. 基于L-ConvNeXt网络的变电站人员安全操作检测方法[J]. 广西师范大学学报（自然科学版）, 2024, 42(1): 102-110.
[8] 黄叶祺, 王明伟, 闫瑞, 雷涛. 基于改进的YOLOv5金刚石线表面质量检测[J]. 广西师范大学学报（自然科学版）, 2023, 41(4): 123-134.
[9] 陈文康, 陆声链, 刘冰浩, 李帼, 刘晓宇, 陈明. 基于改进YOLOv4的果园柑橘检测方法研究[J]. 广西师范大学学报（自然科学版）, 2021, 39(5): 134-146.
[10] 唐熔钗, 伍锡如. 基于改进YOLO-V3网络的百香果实时检测[J]. 广西师范大学学报（自然科学版）, 2020, 38(6): 32-39.
[11] 刘英璇, 伍锡如, 雪刚刚. 基于深度学习的道路交通标志多目标实时检测[J]. 广西师范大学学报（自然科学版）, 2020, 38(2): 96-106.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 广西师范大学学报（自然科学版）编辑部
地址：广西桂林市三里店育才路15号　邮编：541004
电话：0773-5857325　E-mail: gxsdzkb@mailbox.gxnu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发