基于改进YOLOv3-SPP算法的道路车辆检测

doi:10.11959/j.issn.1000-436x.2024046

通信学报 ›› 2024, Vol. 45 ›› Issue (2): 68-78.doi: 10.11959/j.issn.1000-436x.2024046

• 学术论文 • 上一篇

基于改进YOLOv3-SPP算法的道路车辆检测

王涛¹^,², 冯浩¹^,², 秘蓉新³, 李林³, 何振学⁴, 傅奕茗¹^,², 吴姝¹^,²

¹ 北京信息科技大学信息与通信工程学院，北京 100101
² 北京信息科技大学光电测试技术及仪器教育部重点实验室，北京 100192
³ 国家计算机网络应急技术处理协调中心，北京 100029
⁴ 河北农业大学河北省农业大数据重点实验室，河北保定 071001

修回日期:2024-01-17 出版日期:2024-02-01 发布日期:2024-02-01
作者简介:王涛（1986− ），男，陕西渭南人，博士，北京信息科技大学副教授、硕士生导师，主要研究方向为车路协同感知、计算机视觉、雷达数据处理等
冯浩（1996− ），男，河南驻马店人，北京信息科技大学硕士生，主要研究方向为车路协同感知、计算机视觉、雷达数据处理等
秘蓉新（1984− ），女，北京人，国家计算机网络应急技术处理协调中心助理研究员，主要研究方向为人工智能、区块链、网络信息安全等
李林（1986− ），男，安徽巢湖人，博士，国家计算机网络应急技术处理协调中心工程师，主要研究方向为网络信息安全等
何振学（1988− ），男，山东泰安人，博士，河北农业大学副教授、博士生导师，主要研究方向为智能装备、计算机视觉等
傅奕茗（2000− ），男，江苏徐州人，北京信息科技大学硕士生，主要研究方向为智能感知、点云数据处理、激光SLAM等
吴姝（1997− ），女，山西大同人，北京信息科技大学硕士生，主要研究方向为自动驾驶智能感知、激光雷达点云数据处理等
基金资助:
国家自然科学基金资助项目(62001034);国家自然科学基金资助项目(62102130);国家自然科学基金资助项目(6210031547);北京市自然科学基金资助项目(4232004);北京市教育委员会科学研究计划基金资助项目(KM202111232013);河北省自然科学基金资助项目(F2020204003);河北省青年拔尖人才计划基金资助项目(BJ2019008)

Road vehicle detection based on improved YOLOv3-SPP algorithm

Tao WANG¹^,², Hao FENG¹^,², Rongxin MI³, Lin LI³, Zhenxue HE⁴, Yiming FU¹^,², Shu WU¹^,²

¹ School of Information and Communication Engineering, Beijing Information Science ＆Technology University, Beijing 100101, China
² Key Laboratory of Optoelectronic Testing Technology and Instrument, Ministry of Education, Beijing Information Science ＆Technology University, Beijing 100192, China
³ National Computer Network Emergency Response Technical Team/Coordination Center of China, Beijing 100029, China
⁴ Hebei Key Laboratory of Agricultural Big Data, Hebei Agricultural University, Baoding 071001, China

Revised:2024-01-17 Online:2024-02-01 Published:2024-02-01
Supported by:
The National Natural Science Foundation of China(62001034);The National Natural Science Foundation of China(62102130);The National Natural Science Foundation of China(6210031547);The Natural Science Foundation of Beijing(4232004);The Research and Development Program of Beijing Municipal Education Commission(KM202111232013);The Natural Science Foundation of Hebei Province(F2020204003);The Youth Top Talent Program Project of Hebei Province(BJ2019008)

摘要/Abstract

摘要：

针对在城市道路场景下视觉检测车辆时，车辆密集和远处车辆呈现小尺度，导致出现检测精度低或者漏检的问题，提出了一种基于改进的YOLOv3-SPP算法，对激活函数进行优化，以DIOU-NMS Loss作为边界框损失函数，增强网络的表达能力。为提高所提算法对小目标和遮挡目标的特征提取能力，引入空洞卷积模块，增大目标的感受野。实验结果表明，所提算法在检测车辆目标时 mAP 提高了 1.79%，也有效减少了在检测紧密车辆目标时出现的漏检现象。

关键词: 车辆检测, YOLOv3-SPP算法, 激活函数, 空洞卷积, 深度学习

Abstract:

Aiming at the problem of low detection accuracy or missing detection caused by dense vehicles and small scale of distant vehicles in the visual detection of urban road scenes, an improved YOLOv3-SPP algorithm was proposed to optimize the activation function and take DIOU-NMS Loss as the boundary frame loss function to enhance the expression ability of the network.In order to improve the feature extraction ability of the proposed algorithm for small targets and occluding targets, the void convolution module was introduced to increase the receptive field of the target.Based on the experimental results, the proposed algorithm improves the mAP by 1.79% when detecting vehicle targets, and also effectively reduce the missing phenomenon when detecting tight vehicle targets.

Key words: vehicle detection, YOLOv3-SPP algorithm, activation function, atrous convolution, deep learning

中图分类号:

TN92

王涛, 冯浩, 秘蓉新, 李林, 何振学, 傅奕茗, 吴姝. 基于改进YOLOv3-SPP算法的道路车辆检测[J]. 通信学报, 2024, 45(2): 68-78.

Tao WANG, Hao FENG, Rongxin MI, Lin LI, Zhenxue HE, Yiming FU, Shu WU. Road vehicle detection based on improved YOLOv3-SPP algorithm[J]. Journal on Communications, 2024, 45(2): 68-78.

图/表 9

图1

图2

图3

图4

表1

表2

表3

图5

图6

参考文献 27

[1]	刘耀辉 . 基于深度学习的车辆检测研究[D]. 南京:南京邮电大学, 2022.
	LIU Y H . Research on vehicle detection based on deep learning[D]. Nanjing:Nanjing University of Posts and Telecommunications, 2022.
[2]	JABRI S , SAIDALLAH M , BELRHITI E A A ,et al. Moving vehicle detection using Haar-like,LBP and a machine learning adaboost algorithm[C]// Proceedings of IEEE International Conference on Image Processing,Applications and Systems (IPAS). Piscataway:IEEE Press, 2018: 121-124.
[3]	DAVID H , ATHIRA T A . Improving the performance of vehicle detection and verification by Log Gabor filter optimization[C]// Proceedings of Fourth International Conference on Advances in Computing and Communications. Piscataway:IEEE Press, 2014: 50-55.
[4]	GIRSHICK R , DONAHUE J , DARRELL T ,et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition. Piscataway:IEEE Press, 2014: 580-587.
[5]	GU R , BILLINGE S J L , DU Q . A fast two-stage algorithm for non-negative matrix factorization in smoothly varying data[J]. Acta Crystallographica Section A,Foundations and Advances, 2023,79(2): 203-216.
[6]	ZHAO W Q , XU M F , CHENG X F ,et al. An insulator in transmission lines recognition and fault detection model based on improved faster RCNN[J]. IEEE Transactions on Instrumentation and Measurement, 2021,70: 1-8.
[7]	LIU W , ANGUELOV D , ERHAN D ,et al. SSD:single shot multibox detector[C]// European Conference on Computer Vision. Berlin:Springer, 2016: 21-37.
[8]	REDMON J , DIVVALA S , GIRSHICK R ,et al. You only look once:unified,real-time object detection[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2016: 779-788.
[9]	REDMON J , FARHADI A . YOLO9000:better,faster,stronger[C]// Proceedings of IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2017: 6517-6525.
[10]	SHEN J Q , LIU N Z , SUN H ,et al. Vehicle detection in aerial images based on lightweight deep convolutional network and generative adversarial network[J]. IEEE Access, 2019,7: 148119-148130.
[11]	CARION N , MASSA F , SYNNAEVE G ,et al. End-to-end object detection with transformers[C]// European Conference on Computer Vision. Berlin:Springer, 2020: 213-229.
[12]	QIU L R , ZHANG D B , TIAN Y ,et al. Deep learning-based algorithm for vehicle detection in intelligent transportation systems[J]. The Journal of Supercomputing, 2021,77(10): 11083-11098.
[13]	SUN X C , XIA T R , WANG T Y ,et al. A lightweight neural network for vehicle object detection[C]// Proceedings of IEEE 4th International Conference on Civil Aviation Safety and Information Technology (ICCASIT). Piscataway:IEEE Press, 2022: 1103-1106.
[14]	WU Y F , GUAN X R , ZHAO B Y ,et al. Vehicle detection based on adaptive multimodal feature fusion and cross-modal vehicle index using RGB-T images[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2023,16: 8166-8177.
[15]	XU J , LI Z S , DU B W ,et al. ReLUplex made more practical:leaky ReLU[C]// Proceedings of the IEEE Symposium on Computers and Communications (ISCC). Piscataway:IEEE Press, 2020: 1-7.
[16]	CHEN L C , PAPANDREOU G , KOKKINOS I ,et al. DeepLab:semantic image segmentation with deep convolutional nets,atrous convolution,and fully connected CRFs[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018,40(4): 834-848.
[17]	李琦林 . 面向目标检测的边界框回归改进算法研究与实现[D]. 长春:吉林大学, 2023.
	LI Q L . Research and implementation of improved bounding box regression algorithm for target detection[D]. Changchun:Jilin University, 2023.
[18]	JIANG K , ITOH H , ODA M ,et al. Gaussian affinity and GIoU-based loss for perforation detection and localization from colonoscopy videos[J]. International Journal of Computer Assisted Radiology and Surgery, 2023,18(5): 795-805.
[19]	ZHENG Z H , WANG P , LIU W ,et al. Distance-IoU loss:faster and better learning for bounding box regression[J]. arXiv Preprint,arXiv:1911.08287, 2019.
[20]	CAO J , SU Z , YU L Y ,et al. Softmax cross entropy loss with unbiased decision boundary for image classification[C]// Proceedings of Chinese Automation Congress (CAC). Piscataway:IEEE Press, 2018: 2028-2032.
[21]	REZATOFIGHI H , TSOI N , GWAK J ,et al. Generalized intersection over union:a metric and a loss for bounding box regression[C]// Proceedings of IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Piscataway:IEEE Press, 2019: 658-666.
[22]	赵文清, 严海, 邵绪强 . 改进的非极大值抑制算法的目标检测[J]. 中国图象图形学报, 2018,23(11): 1676-1685.
	ZHAO W Q , YAN H , SHAO X Q . Object detection based on improved non-maximum suppression algorithm[J]. Journal of Image and Graphics, 2018,23(11): 1676-1685.
[23]	NI J J , SHEN K , CHEN Y ,et al. An improved SSD-like deep network-based object detection method for indoor scenes[J]. IEEE Transactions on Instrumentation and Measurement, 2023,72: 1-15.
[24]	YANG L , ZHONG J H , ZHANG Y ,et al. An improving faster-RCNN with multi-attention ResNet for small target detection in intelligent autonomous transport with 6G[J]. IEEE Transactions on Intelligent Transportation Systems, 2023,24(7): 7717-7725.
[25]	GAO P , TIAN J W , TAI Y ,et al. Vehicle detection with bottom enhanced RetinaNet in aerial images[C]// Proceedings of IEEE International Geoscience and Remote Sensing Symposium. Piscataway:IEEE Press, 2020: 1173-1176.
[26]	ZHOU S L , XU H , ZHANG G H ,et al. Leveraging deep convolutional neural networks pre-trained on autonomous driving data for vehicle detection from roadside LiDAR data[J]. IEEE Transactions on Intelligent Transportation Systems, 2022,23(11): 22367-22377.
[27]	YANG K L , JIAO Z Y , LIANG J X ,et al. An application case of object detection model based on YOLOv3-SPP model pruning[C]// Proceedings of IEEE International Conference on Artificial Intelligence and Computer Applications (ICAICA). Piscataway:IEEE Press, 2022: 578-582.

名称	相关配置
中央处理器	AMD Ryzen 7 4800H with Radeon Graphics2.90 GHz
内存	DDR4 3200 MHz EEC Reg 32 GB
GPU	NVIDIA GTX1650ti
GPU加速库	CUDA10.2，CUDNN8.7.0
操作系统	Windows 10家庭中文版
数据处理	Python3.9，OpenCV
深度学习框架	PyTorch1.7.1

参数名称	预设值	作用
图片大小	512×512	保持相同的图片预设处理方式，发挥模型的推理效果
迭代次数	1 000	整个数据集迭代次数
批大小	16	对数据集划分，批量处理数据
IOU阈值	0.7	目标检测可信度
置信度阈值	0.45	判断是否存在检测目标

指标	mAP	AP			FPS/Hz
指标	mAP	小汽车	公交车	摩托车	FPS/Hz
FE-CNN	80.31%	82.60%	79.64%	78.69%	25.69
NanoDet	87.59%	88.72%	85.33%	88.72%	17.61
SSD^[23]	76.98%	80.21%	73.28%	77.45%	11.82
Faster-RCNN^[24]	79.63%	84.67%	69.94%	84.28%	29.56
RetinaNet^[25]	81.92%	85.30%	77.26%	83.19%	35.24
Proposed CNN^[26]	84.37%	86.51%	81.92%	84.68%	47.63
传统的YOLOv3-SPP^[27]	88.66%	92.59%	84.17%	89.23%	52.38
改进的YOLOv3-SPP	90.45%	93.47%	85.56%	92.31%	49.54

基于改进YOLOv3-SPP算法的道路车辆检测

Road vehicle detection based on improved YOLOv3-SPP algorithm

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 27

相关文章 15

Metrics

推荐阅读 0

[1]	禹树文, 许威, 姚嘉铖. 面向智能无人通信系统的因果性对抗攻击生成算法[J]. 通信学报, 2024, 45(1): 54-62.
[2]	张学军, 张奉鹤, 盖继扬, 杜晓刚, 周文杰, 蔡特立, 赵博. mVulSniffer：一种多类型源代码漏洞检测方法[J]. 通信学报, 2023, 44(9): 149-160.
[3]	李勉, 李洋, 张纵辉, 史清江. Massive MIMO中通信高效的分布式预编码设计[J]. 通信学报, 2023, 44(8): 37-48.
[4]	王慧娇, 张鑫, 韦永壮, 李灵琛. 基于深度学习的SM4密码算法新型区分器[J]. 通信学报, 2023, 44(7): 171-184.
[5]	李荣鹏, 汪丙炎, 张宏纲, 赵志峰. 知识增强的语义通信接收端设计[J]. 通信学报, 2023, 44(6): 70-76.
[6]	陈东昱, 陈华, 范丽敏, 付一方, 王舰. 基于深度学习的随机性检验策略研究[J]. 通信学报, 2023, 44(6): 23-33.
[7]	马帅, 裴科, 祁华艳, 李航, 曹雯, 王洪梅, 熊海良, 李世银. 基于生成模型的地磁室内高精度定位算法研究[J]. 通信学报, 2023, 44(6): 211-222.
[8]	张昀, 周婧, 黄经纬, 于舒娟, 黄丽亚. 基于深度学习的正交频分复用系统信道估计[J]. 通信学报, 2023, 44(12): 124-133.
[9]	谢刚, 王荃毅, 谢新林, 王健安. 融合多尺度深度卷积的轻量级Transformer交通场景语义分割算法[J]. 通信学报, 2023, 44(10): 213-225.
[10]	杨洁, 董标, 付雪, 王禹, 桂冠. 基于轻量化分布式学习的自动调制分类方法[J]. 通信学报, 2022, 43(7): 134-142.
[11]	杨秀璋, 彭国军, 李子川, 吕杨琦, 刘思德, 李晨光. 基于Bert和BiLSTM-CRF的APT攻击实体识别及对齐研究[J]. 通信学报, 2022, 43(6): 58-70.
[12]	廖育荣, 王海宁, 林存宝, 李阳, 方宇强, 倪淑燕. 基于深度学习的光学遥感图像目标检测研究进展[J]. 通信学报, 2022, 43(5): 190-203.
[13]	廖勇, 王世义. 高速移动环境下基于RM-Net的大规模MIMO CSI反馈算法[J]. 通信学报, 2022, 43(5): 166-176.
[14]	赵增华, 童跃凡, 崔佳洋. 基于域自适应的Wi-Fi指纹设备无关室内定位模型[J]. 通信学报, 2022, 43(4): 143-153.
[15]	廖勇, 程港, 李玉杰. 基于深度展开的大规模MIMO系统CSI反馈算法[J]. 通信学报, 2022, 43(12): 77-88.