基于改进YOLOv5的交通标志检测算法研究
Research on Traffic Sign Detection Algorithm Based on Improved YOLOv5
DOI: 10.12677/CSA.2022.129219, PDF,  被引量    科研立项经费支持
作者: 刘晓倩, 李士心*, 李保胜, 董秀焕, 高雪苹, 陈东园, 张美会：天津职业技术师范大学电子工程学院，天津
关键词: 目标检测；交通标志；YOLOv5算法；轻量化；Object Detection； Traffic Sign； YOLOv5 Algorithm； Lightweight
摘要: 针对交通标志检测网络复杂度高、计算量大、难以进行有效部署的问题，提出一种基于改进YOLOv5的交通标志检测算法。该算法首先使用轻量级卷积神经网络Mobilenev3对原YOLOv5网络的主干特征提取网络进行替换；使用深度可分离卷积替换部分标准卷积，来缩减模型计算量；为保证模型检测精度，引入注意力机制CBAM模块。实验表明，改进后的轻量化网络模型大小只为原YOLOv5模型大小的61%，整理后TT100K数据集上的mAP达到了89.2%，FPS达到了25.1帧/毫秒。本文提出的算法在保证检测精度的前提下，大幅度减少了模型参数量和计算量，并提高了检测速度。
Abstract: Aiming at the problems of high complexity, large amount of computation and difficulty in effective deployment of traffic sign detection network, a traffic sign detection algorithm based on improved YOLOv5 was proposed. Firstly, the lightweight convolutional neural network Mobilenev3 was used to replace the main feature extraction network of the original YOLOv5 network. And the standard convolution was replaced by depth-separable convolution to reduce the computational cost of the model. To ensure the accuracy of model detection accuracy, the attention mechanism CBAM module is introduced. The experimental results show that the size of the improved lightweight network model is only 61% of the size of the original YOLOv5 model, the mAP on TT100K dataset reaches 89.2%, and the FPS reaches 25.1 frames per millisecond. The algorithm proposed in this paper can greatly reduce the number of model parameters and the calculation amount, and improve the detection speed under the premise of ensuring the detection accuracy.
文章引用：刘晓倩, 李士心, 李保胜, 董秀焕, 高雪苹, 陈东园, 张美会. 基于改进YOLOv5的交通标志检测算法研究[J]. 计算机科学与应用, 2022, 12(9): 2161-2168. https://doi.org/10.12677/CSA.2022.129219

参考文献

[1] 陈贺. 中国交通事故死亡状况及酒驾处罚对交通事故死亡率的影响[D]: [硕士学位论文]. 北京: 中国疾病预防控制中心, 2017.
[2] Krizhevskya, Sutskever, I. and Hintong, E. (2017) ImageNet Classification with Deep Convolu-tional Neural Networks. Communications of the ACM, 60, 84-90. [Google Scholar] [CrossRef] 
[3] 曹前. 智能交通中的目标检测技术研究[D]: [硕士学位论文]. 广州: 华南理工大学, 2019.[CrossRef] 
[4] Girshick, R., Donahue, J., Darrell, T., et al. (2014) Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Columbus, 23-28 June 2014, 580-587. [Google Scholar] [CrossRef] 
[5] Girshick, R. (2015) Fast R-CNN. Proceedings of the IEEE Interna-tional Conference on Computer Vision, Santiago, 7-13 December 2015, 1440-1448. [Google Scholar] [CrossRef] 
[6] Girshick, R., Donahue, J., Darrell, T., et al. (2014) Rich Feature Hi-erarchies for Accurate Object Detection and Semantic Segmentation. IEEE Conference on Computer Vision and Pattern Recognition, Columbus, 23-28 June 2014, 580-587. [Google Scholar] [CrossRef] 
[7] He, K., Gkioxari, G., Dollár, P., et al. (2017) Mask R-CNN. Proceedings of the IEEE International Conference on Computer Vision, Ven-ice, 22-29 October 2017, 2961-2969. [Google Scholar] [CrossRef] 
[8] Bochkovskiy, A., Wang, C.Y. and Liao, H.Y.M. (2020) Yolov4: Optimal Speed and Accuracy of Object Detection.
[9] Liu, W., Anguelov, D., Erhan, D., et al. (2016) Ssd: Single Shot Multibox Detector. In: European Conference on Computer Vision, Springer, Cham, 21-37. [Google Scholar] [CrossRef] 
[10] Zhang, J.M., Xie, Z.P., Sun, J., et al. (2020) A Cascaded R-CNN with Multiscale Attention and Imbalanced Samples for Traffic Sign Detection. IEEE Access, 8, 29742-29754. [Google Scholar] [CrossRef] 
[11] Cheng, P., Liu, W., Zhang, Y., et al. (2019) LOCO: Local Context Based Faster R-CNN for Small Traffic Sign Detection. In: International Conference on Multimedia Modeling, Springer, Cham, 329-341. [Google Scholar] [CrossRef] 
[12] Jocher, G., Nishimura, K., Mineeva, T., et al. (2020) YOLOv5.
https://github.com/ultralytics/yolov5
[13] Sandler, M., Howard, A., Zhu, M., et al. (2018) MobileNetV2: Inverted Residuals and Linear Bottlenecks. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Salt Lake City, 18-22 June 2018, 45610-4520. [Google Scholar] [CrossRef] 
[14] Woo, S., Park, J., Lee, J.Y., et al. (2018) CBAM: Convolutional Block Attention Module. Proceedings of the European Conference on Computer Vision (ECCV), Munich, 8-14 September 2018, 3-19. [Google Scholar] [CrossRef] 
[15] Zhu, Z., Liang, D., Zhang, S., et al. (2016) Traffic-Sign Detec-tion and Classification in the Wild. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 27-30 June 2016, 2110-2118. [Google Scholar] [CrossRef]