基于优化YOLO-V4的交通标志检测识别方法

doi:10.11896/jsjkx.220300251

计算机科学 ›› 2022, Vol. 49 ›› Issue (11): 179-184.doi: 10.11896/jsjkx.220300251

• 计算机图形学&多媒体 • 上一篇下一篇

基于优化YOLO-V4的交通标志检测识别方法

潘惠苹^1,2, 王敏琴³, 张福泉⁴

1 广东工商职业技术大学计算机学院广东肇庆 526000
2 菲律宾克里斯汀大学国际学院马尼拉 1004
3 华南理工大学计算机科学与工程学院广州 510000
4 北京理工大学计算机学院北京 100081

收稿日期:2022-03-27 修回日期:2022-05-19 出版日期:2022-11-15 发布日期:2022-11-03
通讯作者: 潘惠苹(15588022@qq.com)
基金资助:
国家自然科学基金面上项目(61871204);福建省科技厅引导性项目(2018H0028)

Traffic Sign Detection and Recognition Method Based on Optimized YOLO-V4

PAN Hui-ping^1,2, WANG Min-qin³, ZHANG Fu-quan⁴

1 College of Computer Science,Guangdong Business and Technology University,Zhaoqing,Guangdong 526000,China
2 Center for International Education,Philippine Christian University,Manila 1004,Philippine
3 College of Computer Science and Engineering,South China University of Science and Engineering,Guangzhou 510000,China
4 College of Computer Science,Beijing Institute of Technology University,Beijing 100081,China

Received:2022-03-27 Revised:2022-05-19 Online:2022-11-15 Published:2022-11-03
About author:PAN Hui-ping,born in 1982,Ph.D,associate professor.Her main research interests include image processing,artificial intelligence and digital media technology.
Supported by:
General Project of the National Natural Science Foundation of China of the(61871204) and Guiding Project of Fujian Provincial Department of Science and Technology(2018H0028).

摘要/Abstract

摘要： 交通标志检测识别是自动驾驶系统的核心功能,为了实时准确地识别交通标志,在YOLO-V4的基础上进行改进,并结合了空间金字塔池化(Spatial Pyramid Pooling,SPP)模块。首先,为了提高分辨率和增大感受野,将原特征图3个尺度的分辨率更改为26×26和52×52;然后,在连接层中添加SPP模块,消除网络对固定尺度的约束,在最大池化层中得到最优特征,改善网络性能。实验中,利用行车记录仪采集各种交通标志图像,与其他优秀方法相比,所提方法取得了更优的性能,其平均检测识别准确度达99.0%,平均检测时间为0.449 s,达到了实时检测的要求。

关键词: 交通标志识别, 感受野, YOLO-V4, 最大池化, 空间金字塔池化, 分辨率

Abstract: Traffic sign detection and recognition is the core function of automatic driving system.In order to identify traffic signs in real time and accurately,a method is improved on the basis of YOLO-V4 and combined with the spatial pyramid pool(SPP) module.Firstly,to increase the resolution and receptive field,the resolution of the three scales of the original feature map is changed to 26×26 and 52×52.Then,SPP module is added to the connection layer to eliminate the constraints of the network on the fixed scale,obtain the optimal characteristics in the maximum pooling layer and improve the network performance.Experiment uses the tachograph to collect various traffic sign images,compared with other excellent methods,the proposed method achieves better performance.The average detection and recognition accuracy of the proposed method is 99.0%,and the average detection time is 0.449 s,which meets the requirements of real-time detection.

Key words: Traffic sign recognition, Receptive field, YOLO-V4, Maxpooling, Spatial pyramid pool, Resolution

中图分类号:

TP391

潘惠苹, 王敏琴, 张福泉. 基于优化YOLO-V4的交通标志检测识别方法[J]. 计算机科学, 2022, 49(11): 179-184. https://doi.org/10.11896/jsjkx.220300251

PAN Hui-ping, WANG Min-qin, ZHANG Fu-quan. Traffic Sign Detection and Recognition Method Based on Optimized YOLO-V4[J]. Computer Science, 2022, 49(11): 179-184. https://doi.org/10.11896/jsjkx.220300251

参考文献

[1]WANG J M,WEI Y,ZHOU Y,et al.Speed limit sign recognition based on LeNet-5 CNN and color feature[J].Computer Science,2021,48(S2):345-350.
[2] ÁLVARO A G,ÁLVAREZ-GARCÍA JUAN A,SORIA-MO-RILLO L M.Evaluation of Deep Neural Networks for traffic sign detection systems[J].Neurocomputing,2018,316(Nov.):332-344.
[3]TAO X,GONG Y,SHI W,et al.Object Detection with ClassAware Region Proposal Network and Focused Attention Objective[J].Pattern Recognition Letters,2018,130(5):353-361.
[4]SHI J,ZHOU Y L,ZHANG Q Z.Service robot item recognition system based on improved mask R-CNN and Kinect[J].Chinese Journal of Science Instrument,2019,40(4):216-228.
[5]DI L,HE R B,LIANG J Z.Road traffic sign recognition algorithm based on possibility clustering and convolution neural network[J].Journal of Nanjing University(Science Edition),2019,55(2):238-250.
[6]DONG N,LIU X Y,WU A G.Detection method for circular traffic signs based on artificial bee colony[J].Journal of Tianjin University(Natural Science and Engineering Technology Edition),2019,52(6):585-593.
[7]ZHANG J,XIE Z,SUN J,et al.A Cascaded R-CNN With Multiscale Attention and Imbalanced Samples for Traffic Sign Detection[J].IEEE Access,2020,51(8):29742-29754.
[8]YAZDAN R,VARSHOSAZ M.Improving traffic sign recognition results in urban areas by overcoming the impact of scale and rotation[J].ISPRS Journal of Photogrammetry and Remote Sensing,2021,171(3):18-35.
[9]LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single ShotMultiBox Detector[C]//European Conference on Computer Vision.2016:21-37.
[10]CHEN P D,HUANG L,XIA Y,et al.Detection and recognition of road traffic signs in UAV images based mask R-CNN[J].Remote Sensing for Land & Resources,2020,32(4):64-70.
[11]PAN W G,LIU B,CHEN Y H,et al.Traffic sign detection and recognition based on YOLO v3[J].Transducer and Microsystem Technologies,2019,38(11):147-150.
[12]CHEN Y L,LU Y R,BAISEN W,et al.Applying deep learning technology to image recognition of traffic signs[C]//International Conference on Technologies and Applications of Artificial Intelligence.Berlin,Germany:2019,IEEE:1-4.
[13]KRIZHEVSKY A,SUTSKEVER I,HINTON G E.ImageNet classification with deep convolutional neural networks[C]//Proceedings of the Advances in Neural Information Processing Systems.Lake Tahoe,USA,2012:1097-1105.
[14]JOSEPH R,ALI F.Yolov3:An incremental improvement[J].arXiv:1804.02767,2018.
[15]WU F,JIN G,GAO M,et al.Helmet detection based on im-proved YOLO v3 deep model[C]//IEEE 16th International Conference on Networking,Sensing and Control.Ban,Canada:IEEE,2019:363-368.
[16]LI W G,YE X,ZHAO Y T,et al.Strip steel surface defect detection based on improved YOLOv3 algorithm[J].Acta Electronica Sinica,2020,48(7):1284-1292.
[17]DEWI C,CHEN R C,YU H.Weight analysis for various prohibitory sign detection and recognition using deep learning[J].Multimedia Tools and Applications,2020,79(43/44):1-19.

相关文章 15

[1]	郭拯危, 付泽文, 李宁, 白澜. 高分辨率斜视聚束SAR回波仿真加速算法研究 Study on Acceleration Algorithm for Raw Data Simulation of High Resolution Squint Spotlight SAR 计算机科学, 2022, 49(8): 178-183. https://doi.org/10.11896/jsjkx.210600066
[2]	戴朝霞, 李锦欣, 张向东, 徐旭, 梅林, 张亮. 基于DNGAN的磁共振图像超分辨率重建算法 Super-resolution Reconstruction of MRI Based on DNGAN 计算机科学, 2022, 49(7): 113-119. https://doi.org/10.11896/jsjkx.210600105
[3]	孟月波, 穆思蓉, 刘光辉, 徐胜军, 韩九强. 基于向量注意力机制GoogLeNet-GMP的行人重识别方法 Person Re-identification Method Based on GoogLeNet-GMP Based on Vector Attention Mechanism 计算机科学, 2022, 49(7): 142-147. https://doi.org/10.11896/jsjkx.210600198
[4]	徐涛, 陈奕仁, 吕宗磊. 基于改进YOLOv3的机坪工作人员反光背心检测研究 Study on Reflective Vest Detection for Apron Workers Based on Improved YOLOv3 Algorithm 计算机科学, 2022, 49(4): 239-246. https://doi.org/10.11896/jsjkx.210200119
[5]	周颖, 常明新, 叶红, 张燕. 基于元迁移的太阳能电池板缺陷图像超分辨率重建方法 Super Resolution Reconstruction Method of Solar Panel Defect Images Based on Meta-transfer 计算机科学, 2022, 49(3): 185-191. https://doi.org/10.11896/jsjkx.210100234
[6]	陈贵强, 何军. 自然场景下遥感图像超分辨率重建算法研究 Study on Super-resolution Reconstruction Algorithm of Remote Sensing Images in Natural Scene 计算机科学, 2022, 49(2): 116-122. https://doi.org/10.11896/jsjkx.210700095
[7]	冷佳旭, 王佳, 莫梦竟成, 陈泰岳, 高新波. 基于深度学习的视频超分辨率重构进展综述 Survey on Video Super-resolution Based on Deep Learning 计算机科学, 2022, 49(2): 123-133. https://doi.org/10.11896/jsjkx.211000007
[8]	周文辉, 石敏, 朱登明, 周军. 基于残差注意力网络的地震数据超分辨率方法 Seismic Data Super-resolution Method Based on Residual Attention Network 计算机科学, 2021, 48(8): 24-31. https://doi.org/10.11896/jsjkx.200900034
[9]	郭琳, 李晨, 陈晨, 赵睿, 范仕霖, 徐星雨. 基于通道注意递归残差网络的图像超分辨率重建 Image Super-resolution Reconstruction Using Recursive ResidualNetwork Based on ChannelAttention 计算机科学, 2021, 48(8): 139-144. https://doi.org/10.11896/jsjkx.200500150
[10]	程松盛, 潘金山. 基于深度学习特征匹配的视频超分辨率方法 Video Super-resolution Method Based on Deep Learning Feature Warping 计算机科学, 2021, 48(7): 184-189. https://doi.org/10.11896/jsjkx.200800224
[11]	崔雯昊, 蒋慕蓉, 杨磊, 傅鹏铭, 朱凌霄. 结合MCycleGAN与RFCNN实现太阳斑点图高分辨重建 Combining MCycleGAN and RFCNN to Realize High Resolution Reconstruction of Solar Speckle Image 计算机科学, 2021, 48(6A): 38-42. https://doi.org/10.11896/jsjkx.201000160
[12]	王伟, 胡涛, 李欣蔚, 沈思婉, 姜小明, 刘峻源. 白细胞图像超分辨率重建研究 Study on Super-resolution Image Reconstruction of Leukocytes 计算机科学, 2021, 48(4): 164-168. https://doi.org/10.11896/jsjkx.200100099
[13]	陆要要, 袁家斌, 何珊, 王天星. 基于超分辨率重建的低质量视频人脸识别方法 Low-quality Video Face Recognition Method Based on Super-resolution Reconstruction 计算机科学, 2021, 48(11A): 295-302. https://doi.org/10.11896/jsjkx.201200159
[14]	谢海平, 李高源, 杨海涛, 赵洪利. 超分辨率重构遥感图像分类研究 Classification Research of Remote Sensing Image Based on Super Resolution Reconstruction 计算机科学, 2021, 48(11A): 424-428. https://doi.org/10.11896/jsjkx.210300132
[15]	刘遵雄, 朱成佳, 黄稷, 蔡体健. 多跳连接残差注意网络的图像超分辨率重建 Image Super-resolution by Residual Attention Network with Multi-skip Connection 计算机科学, 2021, 48(11): 258-267. https://doi.org/10.11896/jsjkx.201000033

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed

基于优化YOLO-V4的交通标志检测识别方法

Traffic Sign Detection and Recognition Method Based on Optimized YOLO-V4

PDF (PC)

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

Metrics

本文评价

推荐阅读 0