计算机科学 ›› 2021, Vol. 48 ›› Issue (1): 89-95.doi: 10.11896/jsjkx.200800034

• 智能化边缘计算* 上一篇    下一篇

L-YOLO:适用于车载边缘计算的实时交通标识检测模型

单美静, 秦龙飞, 张会兵   

  1. 华东政法大学信息科学与技术系 上海 201620
    广西可信软件重点实验室(桂林电子科技大学) 广西 桂林 541004
  • 收稿日期:2020-08-05 修回日期:2020-12-12 出版日期:2021-01-15 发布日期:2021-01-15
  • 通讯作者: 张会兵(zhanghuibing@guet.edu.cn)
  • 作者简介:shanmeijing@ecupl.edu.cn
  • 基金资助:
    国家社科基金一般项目(16BFX085)

L-YOLO:Real Time Traffic Sign Detection Model for Vehicle Edge Computing

SHAN Mei-jing, QIN Long-fei, ZHANG Hui-bing   

  1. Department of Information Science and Technology,East China University of Political Science and Law,Shanghai 201620,China
    Guangxi Key Lab of Trusted Software,Guilin University of Electronic Technology,Guilin,Guangxi 541004,China
  • Received:2020-08-05 Revised:2020-12-12 Online:2021-01-15 Published:2021-01-15
  • About author:SHAN Mei-jing,born in 1979,Ph.D,associate professor.Her main research interests include cybercrime and compu-ter forensics and machine learning.
    ZHANG Hui-bing,born in 1976,Ph.D,associate professor,is a member of China Computer Federation.His main research interests include Internet of Things,edge computing and social computing.
  • Supported by:
    National Social Science Fund of China(16BFX085).

摘要: 在车载边缘计算单元中,由于其硬件设备的资源受限,开发适用于车载边缘计算的轻量级、高效的交通标识检测模型变得越来越迫切。文中提出了一种基于Tiny YOLO改进的轻量级交通标识检测模型,称为L-YOLO。首先,L-YOLO使用部分残差连接来增强轻量级网络的学习能力;其次,为了降低交通标识的误检和漏检,L-YOLO使用高斯损失函数作为边界框的定位损失。在TAD16K交通标识检测数据集上,L-YOLO的参数量为18.8 M,计算量为8.211 BFlops,检测速度为83.3 FPS,同时mAP达到86%。实验结果显示,该算法在保证实时性的同时,还提高了检测精度。

关键词: 车载边缘计算, 目标检测, 交通标识检测, 卷积神经网络, 残差连接, Tiny YOLO

Abstract: In the vehicle edge computing unit,due to the limited resources of its hardware equipment,it becomes more and more urgent to develop a lightweight and efficient traffic sign detection model for vehicle edge computing.This paper proposes a lightweight traffic sign detection model based on Tiny YOLO,which is called L-YOLO.Firstly,L-YOLO uses partial residual connection to enhance the learning ability of lightweight network.Secondly,in order to reduce the false detection and missed detection of traffic signs,L-YOLO uses Gauss loss function as the location loss of boundary box.In the traffic sign detection dataset named TAD16K,the parameter amount of L-YOLO is 18.8M,the calculation amount is 8.211BFlops,the detection speed is 83.3FPS,and the mAP reaches 86%.Experimental results show that the algorithm not only guarantees the real-time performance,but also improves the detection accuracy.

Key words: Vehicle edge computing, Object detection, Traffic sign detection, Convolutional neural network, Residual connection, Tiny YOLO

中图分类号: 

  • TP391
[1] ZHU Z,LIANG D,ZHANG S,et al.Traffic-sign detection and classification in the wild[C]//IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2016:2110-2118.
[2] LI Y,WANG J,XING T,et al.TAD16K:An enhanced benchmark for autonomous driving[C]//IEEE International Confe-rence on Image Processing.IEEE Computer Society,2017:2344-2348.
[3] YANG T,LONG X,SANGAIAH A K,et al.Deep detectionnetwork for real-life traffic sign in vehicular networks[J].Computer Networks,2018,136(8):95-104.
[4] LU Y,JIA M,ZHANG S,et al.Traffic signal detection and classification in street views using an attention model[J].Computational Visual Media,2018,4(3):253-266.
[5] REDMON J,DIVVALA S,GIRSHICK R,et al.You only look once:unified,real-time object detection[C]//IEEE Conference on Computer Vision And Pattern Recognition.IEEE Computer Society,2016:779-788.
[6] REDMON J,FARHADI A.YOLO9000:Better,Faster,Stronger[C]//IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2017:7263-7271.
[7] REDMON J,FARHADI A.YOLOv3:An Incremental Improvement[J].arXiv:1804.02767,2018.
[8] WANG C Y,LIAO H Y M,CHEN P Y,et al.Enriching variety of layer-wise learning information by gradient combination[C]//IEEE/CVF International Conference on Computer Vision Workshop.IEEE Computer Society,2019:2477-2484.
[9] CHOI J,CHUN D,KIM H,et al.Gaussian YOLOv3:An accurate and fast object detector using localization uncertainty for autonomous driving[J].arXiv:1904.04620v2,2019.
[10] GIRSHICK R,DONAHUE J,DARRELL T,et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2014:580-587.
[11] UIJLINGS J R R,VAN DE SANDE K E A,GEVERS T,et al.Selective search for object recognition[J].International Journal of Computer Vision,2013,104(2):154-171.
[12] HE K M,ZHANG X Y,REN S Q,et al.Spatial pyramid pooling in deep convolutional networks for visual Recognition[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2015,37(9):1904-1916.
[13] GIRSHICK,ROSS.Fast R-CNN[C]//IEEE International Conference on Computer vision.IEEE Computer Society,2015:1440-1448.
[14] REN S,HE K,GIRSHICK R,et al.Faster R-CNN:Towards Real-Time Object Detection with Region Proposal Networks[J].IEEE Trans.Pattern Anal.Mach.Intell.,2017,39:1137-1149.
[15] LIU W,ANGUELOV D.SSD:Single Shot MultiBox Detector[C]//The 14th European Conference on Computer Vision.2016:21-37.
[16] SIMONYAN K,ZISSERMAN A.Very deep convolutional networks for large-scale image recognition[C]//3rd International Conference on Learning Representations.2015:1556.
[17] HE K,ZHANG X,REN S,et al.Deep Residual Learning for Image Recognition[C]//IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2016:770-778.
[18] LIN T Y,DOLLAR P,GIRSHICK R,et al.Feature pyramidnetworks for object detection[C]//IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2017:2117-2125.
[19] HE K,GKIOXARI G,DOLLAR P,et al.Mask R-CNN[C]//IEEE international conference on computer vision.IEEE Computer Society,2017:2961-2969.
[20] NEUBECK A,GOOL,Van L J.Efficient Non-Maximum Suppression[C]//International Conference on Pattern Recognition.IEEE Computer Society,2006:850-855.
[21] STALLKAMP J,SCHLIPSING M,SALMEN J,et al.The germantraffic sign recognition benchmark:a multi-class classificationcompetition[C]//The International Joint Conference on Neural Networks.IEEE Computer Society,2011:1453-1460.
[22] GEIGER A,LENZ P,URTASUN R.Are we ready for autonomous driving? The kitti vision benchmark suite [C]//IEEE Conference on Computer Vision and Pattern Recognition.IEEE Computer Society,2012:3354-3361.
[23] REDMON J.Darknet:Open Source Neural Networks in C,2013-2016 [OL].http://pjreddie.com/darknet/.
[24] LIN T Y,MAIRE M,BELONGIE S,et al.Microsoft COCO:Common objectsin context[C]//The 12th European Conference on Computer Vision.2014:740-755.
[1] 杨紫淇, 蔡英, 张皓晨, 范艳芳. 基于负载均衡的VEC服务器联合计算任务卸载方案[J]. 计算机科学, 2021, 48(1): 81-88.
[2] 何彦辉, 吴桂兴, 吴志强. 基于域适应的X光图像的目标检测[J]. 计算机科学, 2021, 48(1): 175-181.
[3] 李亚男, 胡宇佳, 甘伟, 朱敏. 基于深度学习的miRNA靶位点预测研究综述[J]. 计算机科学, 2021, 48(1): 209-216.
[4] 张佳嘉, 张小洪. 多分支卷积神经网络肺结节分类方法及其可解释性[J]. 计算机科学, 2020, 47(9): 129-134.
[5] 朱玲莹, 桑庆兵, 顾婷婷. 基于视差信息的无参考立体图像质量评价[J]. 计算机科学, 2020, 47(9): 150-156.
[6] 崔彤彤, 王桂玲, 高晶. 基于1DCNN-LSTM的船舶轨迹分类方法[J]. 计算机科学, 2020, 47(9): 175-184.
[7] 齐少华, 徐和根, 万友文, 付豪. 动态环境下的语义地图构建[J]. 计算机科学, 2020, 47(9): 198-203.
[8] 刘海潮, 王莉. 基于深度图卷积胶囊网络的图分类模型[J]. 计算机科学, 2020, 47(9): 219-225.
[9] 姚兰, 赵永恒, 施雨晴, 于明鹤. 一种基于视频分析的高速公路交通异常事件检测算法[J]. 计算机科学, 2020, 47(8): 208-212.
[10] 黄金星, 潘翔, 郑河荣. 基于残差连接的场景文本识别端到端网络结构优化[J]. 计算机科学, 2020, 47(8): 221-226.
[11] 梁正友, 何景琳, 孙宇. 一种用于微表情自动识别的三维卷积神经网络进化方法[J]. 计算机科学, 2020, 47(8): 227-232.
[12] 孙彦丽, 叶炯耀. 基于剪枝与量化的卷积神经网络压缩方法[J]. 计算机科学, 2020, 47(8): 261-266.
[13] 程哲, 白茜, 张浩, 王世普, 梁宇. 使用深层卷积神经网络提高Hi-C 数据分辨率[J]. 计算机科学, 2020, 47(6A): 70-74.
[14] 赫磊, 邵展鹏, 张剑华, 周小龙. 基于深度学习的行为识别算法综述[J]. 计算机科学, 2020, 47(6A): 139-147.
[15] 孙正, 王新宇. 深度学习在光声成像中的应用现状[J]. 计算机科学, 2020, 47(6A): 148-152.
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
[1] 郭俊霞,郭仁飞,许南山,赵瑞莲. 基于Session的Web应用软件EFSM模型构建方法研究[J]. 计算机科学, 2018, 45(4): 203 -207 .
[2] 项英倬, 谭菊仙, 韩杰思, 石浩. 图匹配技术研究[J]. 计算机科学, 2018, 45(6): 27 -31 .
[3] 金瑞, 刘作学. 一种采用时隙对准方式的TDMA自组网同步协议[J]. 计算机科学, 2018, 45(6): 84 -88 .
[4] 季海娟, 周从华, 刘志锋. 一种基于始末距离的时间序列符号聚合近似表示方法[J]. 计算机科学, 2018, 45(6): 216 -221 .
[5] 陈福才, 李思豪, 张建朋, 黄瑞阳. 基于标签关系改进的多标签特征选择算法[J]. 计算机科学, 2018, 45(6): 228 -234 .
[6] 陈嵘, 李鹏, 黄勇. 基于多特征融合的运动阴影去除算法[J]. 计算机科学, 2018, 45(6): 291 -295 .
[7] 田晓燕,魏娜,范泽铭,张锁良. RA码交织器的研究与设计[J]. 计算机科学, 2018, 45(5): 79 -82 .
[8] 敖权,陆慧梅,向勇,曹睿东. 基于QEMU的Linux应用异常通信行为分析[J]. 计算机科学, 2018, 45(5): 89 -96 .
[9] 潘明明,李丁丁,汤庸,刘海. 一种基于中间件的异构数据库融合访问方法及系统[J]. 计算机科学, 2018, 45(5): 163 -167 .
[10] 徐凤生,于秀清,史开泉. P-数据模型与数据的智能获取[J]. 计算机科学, 2018, 45(5): 176 -179 .