基于Libra R-CNN改进的交通标志检测算法

doi:10.3901/JME.2021.22.255

机械工程学报 ›› 2021, Vol. 57 ›› Issue (22): 255-265.doi: 10.3901/JME.2021.22.255

扫码分享

基于Libra R-CNN改进的交通标志检测算法

赵子婧¹, 刘宏哲¹, 曹东璞²

1. 北京联合大学北京市信息服务工程重点实验室北京 100101;
2. 滑铁卢大学认知自动驾驶实验室滑铁卢 ON N2L 3G1 加拿大

收稿日期:2020-11-10 修回日期:2021-06-26 出版日期:2021-11-20 发布日期:2022-02-28
通讯作者: 刘宏哲(通信作者),女,1971年出生,博士,教授,硕士研究生导师。主要研究方向为人工智能、视觉智能、认知计算和视觉计算。E-mail:liuhongzhe@buu.edu.cn
作者简介:赵子婧,女,1996年出生。主要研究方向为计算机视觉、无人驾驶。E-mail:181083520409@buu.edu.cn;曹东璞,男,1978年出生,博士,副教授。主要研究方向为车辆控制和智能化、自动驾驶、平行驾驶。E-mail:dongpu.cao@uwaterloo.ca
基金资助:
国家自然科学基金（61871039，61906017，61802019）、北京市教委（KM202111417001，KM201911417001）、视觉智能协同创新中心（CYXC2011）、北京联合大学学术研究（ZB10202003，ZK80202001，XP202015）和北京联合大学研究生资助项目。

Improved Traffic Sign Detection Algorithm Based on Libra R-CNN

ZHAO Zijing¹, LIU Hongzhe¹, CAO Dongpu²

1. Beijing Key Laboratory of Information Service Engineering, Beijing Union University, Beijing 100101;
2. Waterloo Cognitive Autonomous Driving (CogDrive) Lab, University of Waterloo, Waterloo ON N2L 3G1, Canada

Received:2020-11-10 Revised:2021-06-26 Online:2021-11-20 Published:2022-02-28

摘要/Abstract

摘要： 随着人工智能领域的快速发展，深度学习在无人驾驶领域中的应用逐渐成熟，但是其中交通标志检测任务作为难点问题仍有很大的改进空间。城市道路下的交通标志检测具有环境复杂、小目标多、目标种类多且数量不平衡的特点，针对这些问题，提出基于Libra R-CNN进行改进的方案。Libra R-CNN目标检测网络是基于平衡提出的，能够较好应对目标种类多及数量不平衡问题，在Libra R-CNN网络的锚框提取样本阶段，使用GA-RPN生成锚框，从而在训练期间产生更精确、更多样化的样本，减少背景影响和小目标不好定位的问题，提高检测准确率。该方法通过试验验证了有效性。试验是在MS COCO 2017和交通标志数据集上进行的。改进后的Libra R-CNN的mAP提高了超2.7个百分点。试验结果表明，改进后的网络相比原有的目标检测网络性能有了显著提升。

关键词: 计算机视觉, 深度学习, 目标检测, 交通标志检测, 改进的LibraR-CNN, GA-RPN

Abstract: In order to study the potential benefits and constraints of controlling vehicle braking in reducing ground related injury, 139 real world pedestrian-vehicle accidents are selected from the pre-accumulated accident database firstly, and then all selected accidents are reconstructed by PC-Crash, after that a vehicle braking method is applied to each reconstructed cases, finally the injury in various parts of human body, collision location of the head-car and head- ground, and temporal/spatial constraints during braking control are collected. Results show that the vehicle full braking after the impact does not have a significant influence on pedestrian injury in real world accidents; through controlling vehicle braking, the ground related head/pelvis injury can be reduced significantly but the vehicle related injury will not increase, and the coincidence rate of the collision position of head-car and head-ground can be reduced; but controlling vehicle braking requires the vehicle to make a judgment in a short time to correctly control the vehicle and 8.4% of cases do not have enough space for the vehicle to carry out a braking control.

Key words: computer vision, deep learning, target detection, traffic sign detection, improved Libra R-CNN, GA-RPN

中图分类号:

U495

赵子婧, 刘宏哲, 曹东璞. 基于Libra R-CNN改进的交通标志检测算法[J]. 机械工程学报, 2021, 57(22): 255-265.

ZHAO Zijing, LIU Hongzhe, CAO Dongpu. Improved Traffic Sign Detection Algorithm Based on Libra R-CNN[J]. Journal of Mechanical Engineering, 2021, 57(22): 255-265.

参考文献

[1] HAN Cen, GAO Guangyu, ZHANG Yu. Real-time small traffic sign detection with revised faster-RCNN[J]. Multimedia Tools and Applications, 2019, 78(10):13263-13278.
[2] SONG Shijin, QUE Zhiqiang, HOU Junjie, et al. An efficient convolutional neural network for small traffic sign detection[J]. Journal of Systems Architecture, 2019, 97:269-277.
[3] WALI SAFAT B, ABDULLAH MAJID A, HANNANMAHAMMAD A, et al. Vision-based traffic sign detection and recognition systems:Current trends and challenges[J]. Sensors(Basel, Switzerland), 2019, 19(9):2093.
[4] 董朋林. 街道实景中的交通标志检测与识别[J]. 现代计算机, 2019(36):37-43. DONG Penglin. Detection and recognition of traffic signs in the real scene of street[J]. Modern Computer, 2019(36):37-43.
[5] PANG J, CHEN K, SHI J, et al. Libra R-CNN:Towards balanced learning for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2019:821-830.
[6] KRIZHEVSKY A, SUTSKEVER I, HINTON G E. Imagenet classification with deep convolutional neural networks[C]//Advances in Neural Information Processing Systems, 2012:1097-1105.
[7] VISHWAKARMA S, AGRAWAL A. A survey on activity recognition and behavior understanding in video surveillance[J]. Visual Computer, 2013, 29(10):983-1009.
[8] 周晓彦, 王珂, 李凌燕. 基于深度学习的目标检测算法综述[J]. 电子测量技术, 2017, 40(11):89-93. ZHOU Xiaoyan, WANG Ke, LI Lingyan. Review of object detection based on deep learning[J]. Electronic Measurement Technology, 2017, 40(11):89-93.
[9] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once:Unified, real-time object detection[C]//Proceedings of the IEEE conference on computer vision and pattern recognition, 2016:779-788.
[10] LIU W, ANGUELOV D, ERHAN D, et al. SSD:Single shot multibox detector[C]//European Conference on Computer Vision. Springer, Cham, 2016:21-37.
[11] GIRSHICK R. Fast R-CNN[C]//Proceedings of the IEEEInternational Conference on Computer Vision, 2015:1440-1448.
[12] REN S, HE K, GIRSHICK R, et al. Faster R-CNN:Towards real-time object detection with region proposal networks[C]//Advances in Neural Information Processing Systems, 2015:91-99.
[13] HE K, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision, 2017:2961-2969.
[14] CAI Z, VASCONCELOS N. Cascade R-CNN:Delving into high quality object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018:6154-6162.
[15] REDMON J, FARHADI A. YOLO9000:Better, faster, stronger[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017:7263-7271.
[16] REDMON J, FARHADI A. YOLOv3:An incremental improvement[R]. ar Xiv preprint ar Xiv:1804. 02767, 2018.
[17] YANG B, YAN J, LEI Z, et al. Craft objects from images[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016:6043-6051.
[18] HUANG S W, LIN C T, CHEN S P, et al. Auggan:Cross domain adaptation with GAN-based data augmentation[C]//Proceedings of the European Conference on Computer Vision (ECCV), 2018:718-731.
[19] WANG X, SHRIVASTAVA A, GUPTA A. A-fast-RCNN:Hard positive generation via adversary for object detection[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017:2606-2615.
[20] 单义, 杨金福, 武随烁, 等. 基于跳跃连接金字塔模型的小目标检测[J]. 智能系统学报, 2019, 14(6):1144-1151. SHAN Yi, YANG Jinfu, WU Suishuo, et al. Skip feature pyramid network with a global receptive field for small object detection[J]. CAAI Transactions on Intelligent Systems, 2019, 14(6):1144-1151.
[21] WANG J, CHEN K, YANG S, et al. Region proposal by guided anchoring[C]//Proceedings of the IEEEConference on Computer Vision and Pattern Recognition, 2019:2965-2974.
[22] REZATOFIGHI H, TSOI N, GWAK J Y, et al. Generalized intersection over union:A metric and a loss for bounding box regression[C]//Proceedings of the IEEEConference on Computer Vision and Pattern Recognition, 2019:658-666.
[23] WANG X, GIRSHICK R, GUPTA A, et al. Non-local neural networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018:7794-7803.
[24] HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition[C]//Proceedings of the IEEEConference on Computer Vision and Pattern Recognition, 2016:770-778.

基于Libra R-CNN改进的交通标志检测算法

Improved Traffic Sign Detection Algorithm Based on Libra R-CNN

PDF

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 15

编辑推荐

Metrics

本文评价

[1]	李春凯, 王嘉昕, 石玗, 代悦. GTAW熔池激光条纹动态行为与熔透预测方法研究[J]. 机械工程学报, 2024, 60(6): 236-244.
[2]	赵德望, 姜超, 赵延广, 杨文平, 樊俊铃. 铝锂合金疲劳性能及预测研究-基于试验和“浅层”+“深度”混合神经网络方法[J]. 机械工程学报, 2024, 60(16): 190-199.
[3]	付玲, 佘玲娟, 颜镀镭, 张鹏, 龙湘云. 基于内嵌物理信息与注意力机制BiLSTM神经网络的臂架系统疲劳损伤预测模型[J]. 机械工程学报, 2024, 60(13): 205-215.
[4]	陈钱, 陈康康, 董兴建, 皇甫一樊, 彭志科, 孟光. 一种面向机械设备故障诊断的可解释卷积神经网络[J]. 机械工程学报, 2024, 60(12): 65-76.
[5]	刘岳开, 王天杨, 褚福磊. 基于低延迟可解释性深度学习的复杂旋转机械关键部件知识嵌入与诊断方法研究[J]. 机械工程学报, 2024, 60(12): 107-115.
[6]	赖旭伟, 丁昆, 张楷, 黄锋飞, 郑庆, 李致萱, 丁国富. 基于可解释物理引导空间注意力改进的跨工艺参数立铣刀磨损辨识[J]. 机械工程学报, 2024, 60(12): 147-157.
[7]	林昕, 毛杨坤, 傅盈西, 朱锟鹏. 基于图像语义分割与图卷积的选区激光熔融成形过程羽流运动特征分析[J]. 机械工程学报, 2024, 60(12): 168-182.
[8]	邵海东, 肖一鸣, 邓乾旺, 任颖莹, 韩特. 基于不确定性感知网络的可信机械故障诊断[J]. 机械工程学报, 2024, 60(12): 194-206.
[9]	林晨, 何智成, 黄怡菲, 林智桂, 付广, 黄晋. 多级参数融合网络的驾驶场景目标检测方法研究[J]. 机械工程学报, 2024, 60(10): 64-75.
[10]	李玖法, 邹博文, 任玥. 考虑车-路交互作用的车辆轨迹预测算法研究[J]. 机械工程学报, 2024, 60(10): 76-85.
[11]	许思远, 张卫东, 毛玉明, 牛斌. 数据驱动的指定变形非均质多胞结构优化设计[J]. 机械工程学报, 2024, 60(1): 85-95.
[12]	毛杨坤, 段现银, 林昕, 傅盈西, 朱锟鹏. 基于目标检测的选区激光熔融成形过程熔池与飞溅监测[J]. 机械工程学报, 2023, 59(9): 335-348.
[13]	何赟泽, 李响, 王洪金, 侯岳骏, 张帆, 牟欣颖, 刘浩, 程豪, 李时华, 李杰. 基于可见光和热成像的风机叶片全周期无损检测综述[J]. 机械工程学报, 2023, 59(6): 32-45.
[14]	彭湃, 耿可可, 王子威, 柳智超, 殷国栋. 智能汽车环境感知方法综述[J]. 机械工程学报, 2023, 59(20): 281-303.
[15]	于宗营, 沈功田, 赵章焰, 吴占稳, 刘渊. 计算机视觉结构健康监测的研究进展及在大型游乐设施上的应用展望[J]. 机械工程学报, 2023, 59(16): 167-181.