基于不可分小波和改进YOLOv8的交通标志检测算法
Traffic Sign Detection Algorithm Based on Non-Separable Wavelets and Improved YOLOv8
DOI: 10.12677/jisp.2024.132016, PDF,   
作者: 范阳旭：湖北大学计算机与信息工程学院，湖北 武汉
关键词: 交通标志检测；不可分小波；YOLOv8；小目标检测；SPD-Conv；Traffic Sign Detection； Non-Separable Wavelet； YOLOv8； Small Object Detection； SPD-Conv
摘要: 在现代智能交通系统中，高效且准确的交通标志检测对于辅助驾驶和自动驾驶系统具有重要意义。针对背景复杂的道路场景中交通标志尺寸小导致的识别精度低，漏检等问题，提出了一种基于不可分小波和改进YOLOv8的交通标志检测算法。首先，采用不可分小波处理输入图像，有效提取高频信息以增强图像的细节表现，提高模型的鲁棒性。其次，引入针对小目标的检测层，取代原始模型中的大目标检测层，优化网络结构，从而显著提升了小目标的检测性能。接着，将网络中的跨步卷积替换成SPD-Conv，有效减少特征信息的丢失。最后，采用WIoU损失函数代替原有的损失函数。在TT100K数据集上进行训练，实验结果显示，改进后的算法相较于YOLOv8在精确率及mAP@0.5上，分别提升了9.7%和11.5%，性能明显优于原始算法。
Abstract: In modern intelligent transportation systems, efficient and accurate traffic sign detection is of great significance for assisted driving and autonomous driving systems. Aiming at the problems of low recognition accuracy and missed detection caused by the small size of traffic signs in road scenes with complex backgrounds, a traffic sign detection algorithm based on indivisible wavelets and improved YOLOv8 is proposed. Firstly, the input image is processed with Non-Separable wavelets to effectively extract high-frequency information in order to enhance the detail performance of the image and improve the robustness of the model. Second, a detection layer for small targets is introduced to replace the large target detection layer in the original model, and the network structure is optimized, thus significantly improving the detection performance of small targets. Next, the stepwise convolution in the network is replaced with SPD-Conv to effectively reduce the loss of feature information. Finally, the WIoU loss function is used instead of the original loss function. Trained on the TT100K dataset, the experimental results show that the improved algorithm improves 9.7% and 11.5% compared to YOLOv8 in terms of accuracy and mAP@0.5, respectively, and the performance is significantly better than the original algorithm.
文章引用：范阳旭. 基于不可分小波和改进YOLOv8的交通标志检测算法[J]. 图像与信号处理, 2024, 13(2): 179-189. https://doi.org/10.12677/jisp.2024.132016

参考文献

[1] 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] 
[2] Girshick, R. (2015) Fast R-CNN. Proceedings of the IEEE International Conference on Computer Vision, Santiago, 7-13 December 2015, 1440-1448. [Google Scholar] [CrossRef] 
[3] Ren, S., He, K., Girshick, R., et al. (2017) Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 1137-1149. [Google Scholar] [CrossRef] 
[4] Liu, W., Anguelov, D., Erhan, D., et al. (2016) Ssd: Single Shot Multibox Detector. In: Leibe, B., Matas, J., Sebe, N. and Welling, M., Eds., Computer Vision—ECCV 2016, Lecture Notes in Computer Science, Vol. 9905, Springer, Cham, 21-37. [Google Scholar] [CrossRef] 
[5] Redmon, J., Divvala, S., Girshick, R., et al. (2016) You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, 27-30 June 2016, 779-788. [Google Scholar] [CrossRef] 
[6] Redmon, J. and Farhadi, A. (2017) YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, 21-26 July 2017, 7263-7271. [Google Scholar] [CrossRef] 
[7] Redmon, J. and Farhadi, A. (2018) Yolov3: An Incremental Improvement. arXiv preprint arXiv:1804.02767.
[8] Bochkovskiy, A., Wang, C.Y. and Liao, H.Y.M. (2020) Yolov4: Optimal Speed and Accuracy of Object Detection. arXiv preprint arXiv:2004.10934.
[9] Wu, Y., Li, Z., Chen, Y., et al. (2020) Real-Time Traffic Sign Detection and Classification towards Real Traffic Scene. Multimedia Tools and Applications, 79, 18201-18219. [Google Scholar] [CrossRef] 
[10] Zhang, H., Qin, L., Li, J., et al. (2020) Real-Time Detection Method for Small Traffic Signs Based on Yolov3. Ieee Access, 8, 64145-64156. [Google Scholar] [CrossRef] 
[11] Yao, Y., Han, L., Du, C., et al. (2022) Traffic Sign Detection Algorithm Based on Improved YOLOv4-Tiny. Signal Processing: Image Communication, 107, Article 116783. [Google Scholar] [CrossRef] 
[12] Wang, J., Chen, Y., Dong, Z., et al. (2023) Improved YOLOv5 Network for Real-Time Multi-Scale Traffic Sign Detection. Neural Computing and Applications, 35, 7853-7865. [Google Scholar] [CrossRef] 
[13] Chen, Q., Micchelli, C.A., Peng, S., et al. (2003) Multivariate Filter Banks Having Matrix Factorizations. SIAM Journal on Matrix Analysis and Applications, 25, 517-531. [Google Scholar] [CrossRef] 
[14] 刘斌, 彭嘉雄. 基于四通道不可分加性小波的多光谱图像融合[J]. 计算机学报, 2009, 32(2): 350-356.
[15] Sunkara, R. and Luo, T. (2022) No More Strided Convolutions or Pooling: A New CNN Building Block for Low-Resolution Images and Small Objects. In: Amini, M.R., Canu, S., Fischer, A., Guns, T., Kralj Novak, P. and Tsoumakas, G., Eds., Machine Learning and Knowledge Discovery in Databases, Lecture Notes in Computer Science, Vol. 13715, Springer, Cham, 443-459. [Google Scholar] [CrossRef] 
[16] Tong, Z., Chen, Y., Xu, Z., et al. (2023) Wise-IoU: Bounding Box Regression Loss with Dynamic Focusing Mechanism. arXiv preprint arXiv:2301.10051.