摘 要：在智能交通系统中，车辆目标检测有广泛应用。为了提高车辆目标检测性能，采用基于 FPN 的 YOLOv3 算法进行车辆多目标检测，并且通过添加注意力机制模块进行网络优化，提出了一种基于空间注意力机制 SAM 的 YOLOv3 车辆多目标检测优化算法，并在所构造的车辆多目标数据集上对提出的算法进行了验证，证明其对车辆多目标检测的优势。实验表明，优化后的检测算法相比原检测算法模型参数量降低了 55.36%，mAP 值提升了 1.15%，优于原检测算法。
关键词：车辆目标检测；注意力机制；YOLOv3；SAM
DOI:10.19850/j.cnki.2096-4706.2021.06.026
基金项目：中原科技创新领军人才（214200 510013）；河南省高校重点科研项目（21A51001 6）；留学人员科研资助和创业启动项目（HRSS20 21[36]）
中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2021）06-0103-04
Research on Vehicle Target Detection Algorithm Based on Attention Mechanism
LUO Jianchen，YANG Lei
（School of Electronic and Information，Zhongyuan University of Technology，Zhengzhou 450007，China）
Abstract：Vehicle target detection is widely used in intelligent transportation system. In order to improve vehicle target detection performance，the FPN-based YOLOv3 algorithm is used for vehicle multi-target detection，and the attention mechanism module is added to optimize the network. An optimized YOLOv3 vehicle multi-target detection algorithm based on spatial attention mechanism（SAM） is proposed. The proposed algorithm is verified on the constructed vehicle multi-target dataset，which proves its advantage in multitarget vehicle detection. The experimental results show that compared with the original detection algorithm，the model parameters of the optimized detection algorithm are reduced by 55.36%，and the mAP value is increased by 1.15%，which is better than the original detection algorithm.
Keywords：vehicle target detection；attention mechanism；YOLOv3；SAM
参考文献
[1] 肖雨晴，杨慧敏.目标检测算法在交通场景中应用综述 [J]. 计算机工程与应用，2021，57（6）：30-41.
[2] JIAO L C，ZHANG F，LIU F，et al. A Survey of Deep Learningbased Object Detection [J].IEEE Access，2019，7：128837-128868.
[3] LIN T Y，DOLLAR P，GIRSHICK R，et al. Feature Pyramid Networks for Object Detection [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition （CVPR）. Honolulu：IEEE，2017：936-994.
[4] HU J，SHEN L，SAMUEL A，et al. Squeeze-and-Excitation Networks [J].IEEE Transactions on Pattern Analysis and Machine Intelligence，2020，42（8）：2011-2023.
[5] WOO S，PARK J，LEE J Y，et al. CBAM：Convolutional Block Attention Module [C]//ECCV：European Conference on Computer Vision.Munich：Springer，2018：3-19.
[6] OpenITS. OpenData V11.0- 车辆重识别数据集 VRID [EB/OL].[2021-02-11].https://www.openits.cn/openData4/ 748.jhtml.
[7] DONG Z，WU Y W，PEI M T，et al. Vehicle Type Classification Using a Semisupervised Convolutional Neural Network [J].IEEE Transactions on Intelligent Transportation Systems，2015，16（4）：2247-2256.
[8] XU Z B，YANG W，MENG A J，et al. Towards End-to-End License Plate Detection and Recognition：A Large Dataset and Baseline [C]//ECCV：European Conference on Computer Vision.Munich： Springer，2018：261-277.
作者简介：罗建晨（1993—），男，汉族，河南信阳人，硕士 研究生在读，主要研究方向：计算机视觉，深度学习；杨蕾（1979 —），女，回族，河南洛阳人，教授，博士，主要研究方向：图像 处理，计算机视觉。