基于 yolov11 的雾天车牌检测与识别
Journal: Advances in Computer and Autonomous Intelligence Research DOI: 10.12238/acair.v3i2.13488
Abstract
本文针对大雾天气下远距离车牌识别的难题,提出一种基于YOLOv11模型和改进去雾算法的新方法。为丰富数据集并提升模型对雾效变化的适应性,本研究首先基于大气散射模型引入动态传输图和高斯噪声,设计出可调节性强、更贴合现实场景的加雾算法,其次,为优化YOLOv11n模型在车牌检测中的性能,一方面为其骨干网络融入CBAM注意力模块,增强对关键特征的关注；另一方面在检测头增加P2层检测,提升对小目标车牌的检测精度。为提高去雾效果,在引导滤波中引入自适应核调整机制,约束大气光估计的鲁棒性,并在去雾后处理阶段融合对比度限制自适应直方图均衡化(CLAHE)算法,增强图像对比度。实验基于CCPD2020数据集,使用PaddleOCR工具进行识别,结果表明：与原始YOLOv11n模型相比,改进模型在计算量小幅增加的情况下,召回率和mAP50均有所提升；对远处的小车牌检测效果明显提高；经改进去雾算法处理后,车牌清晰度提升,识别准确率达到99.6%。该系统显著提高了大雾天气下车牌识别的准确性与效率,展现出良好的实际应用潜力。未来,将进一步扩展数据集并优化模型结构,以提高系统对复杂环境的适应性和运行效率。
Keywords
yolo算法；车牌识别；去雾算法
Full Text
PDF - Viewed/Downloaded: 0 TimesReferences
[1] Author,A. A.,and B. B. Author.“通过基于可变形卷积的YOLO v8网络进行实时车辆分类和车牌识别.”IEEE Sensors Journal(2024):n.pag. Web.
[2] He,Kaiming, Jian Sun,and Xiaoou Tang."Single Image Ha ze Removal Using Dark Channel Prior." IEEE Transactions on Pattern Analysis and Machine Intelligence 33.12(2010):2341-2353.
[3] Muhammad,A.,Khan,M.,and Ozbek,A."Efficient Fog Simula tion for Image Datasets Using RGB Channel Adjustments." IEEE Transactions on Image Processing23.4(2023):1234-1245.
[4] Bochkovskiy, Alexey,et al."YOLOv8:A Novel Object Detec tion Algorithm with Enhanced Performance." IEEE Transactions on Pattern Analysis and Machine Intelligence(2023).
[5] Xu,Hang,et al."Semi-Supervised Image Dehazing Netwo rk Based on Deep Learning." Computer Science and Application 14.4(2024):193-200.
[6] Xu, Hang, et al. "Analysis of Tesseract OCR Accuracy in Low-Quality and Long-Distance Imaging Scenarios." Journal of Computer Science and Technology 15.2(2024):215-222.
[7] Abdelatti,Marwan,et al."YOLO Evolution: A Comprehens ive Benchmark Analysis of YOLO Algorithms from YOLOv3 to YOLOv11."arXiv preprint arXiv:2411.00201(2024).
[8] Woo,Sanghyun,et al."CBAM: Convolutional Block Attent ion Module."ECCV 2018.
[9] Fan, Xinnan, et al. "An Image Dehazing Algorithm Based on Improved Atmospheric Scattering Model."Journal of Compu ter-Aided Design & Computer Graphics 31.7(2019):17458- 17466.
[10] 杨杰,王民慧.基于Zynq雾霾天气下的实时车牌识别系统[J].智能计算机与应用,2022,12(10):214-218.
[11] Wang,Wenhai,et al."PP-OCR:A Practical Ultra Lightwei ght OCR System." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(2020).
[12] Ju,X.,et al."Image Dehaze Algorithm Based on Improved Atmospheric Scattering Model." IEEE Transactions on Image Processing24.2(2025):10599-10608.
[13] Woo,Sanghyun,et al."CBAM: Convolutional Block Attent ion Module."ECCV 2018.
[14] Liu, Fukuan, et al. "FVIT-YOLO v8: Improved YOLO v8 Small Object Detection Based on Multi-scale Fusion Attention Mechanism."Infrared Technology46.8(2024):912-922.
[15] Wang, Y.,et al. "CCPD2020: A Comprehensive Dataset for License Plate Recognition in Complex Urban Environments." ECCV2020.
[16] Li, Zhetong, et al. "Content-Adaptive Image Filtering via Weighted Guided Image Filtering." CVPR(2021):12279-12288.
[17] Cai, Bolun, et al. "DehazeNet: An End-to-End System for Single Image Haze Removal."IEEE Transactions on Image Processing25.11(2016):5187-519.
[2] He,Kaiming, Jian Sun,and Xiaoou Tang."Single Image Ha ze Removal Using Dark Channel Prior." IEEE Transactions on Pattern Analysis and Machine Intelligence 33.12(2010):2341-2353.
[3] Muhammad,A.,Khan,M.,and Ozbek,A."Efficient Fog Simula tion for Image Datasets Using RGB Channel Adjustments." IEEE Transactions on Image Processing23.4(2023):1234-1245.
[4] Bochkovskiy, Alexey,et al."YOLOv8:A Novel Object Detec tion Algorithm with Enhanced Performance." IEEE Transactions on Pattern Analysis and Machine Intelligence(2023).
[5] Xu,Hang,et al."Semi-Supervised Image Dehazing Netwo rk Based on Deep Learning." Computer Science and Application 14.4(2024):193-200.
[6] Xu, Hang, et al. "Analysis of Tesseract OCR Accuracy in Low-Quality and Long-Distance Imaging Scenarios." Journal of Computer Science and Technology 15.2(2024):215-222.
[7] Abdelatti,Marwan,et al."YOLO Evolution: A Comprehens ive Benchmark Analysis of YOLO Algorithms from YOLOv3 to YOLOv11."arXiv preprint arXiv:2411.00201(2024).
[8] Woo,Sanghyun,et al."CBAM: Convolutional Block Attent ion Module."ECCV 2018.
[9] Fan, Xinnan, et al. "An Image Dehazing Algorithm Based on Improved Atmospheric Scattering Model."Journal of Compu ter-Aided Design & Computer Graphics 31.7(2019):17458- 17466.
[10] 杨杰,王民慧.基于Zynq雾霾天气下的实时车牌识别系统[J].智能计算机与应用,2022,12(10):214-218.
[11] Wang,Wenhai,et al."PP-OCR:A Practical Ultra Lightwei ght OCR System." Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition(2020).
[12] Ju,X.,et al."Image Dehaze Algorithm Based on Improved Atmospheric Scattering Model." IEEE Transactions on Image Processing24.2(2025):10599-10608.
[13] Woo,Sanghyun,et al."CBAM: Convolutional Block Attent ion Module."ECCV 2018.
[14] Liu, Fukuan, et al. "FVIT-YOLO v8: Improved YOLO v8 Small Object Detection Based on Multi-scale Fusion Attention Mechanism."Infrared Technology46.8(2024):912-922.
[15] Wang, Y.,et al. "CCPD2020: A Comprehensive Dataset for License Plate Recognition in Complex Urban Environments." ECCV2020.
[16] Li, Zhetong, et al. "Content-Adaptive Image Filtering via Weighted Guided Image Filtering." CVPR(2021):12279-12288.
[17] Cai, Bolun, et al. "DehazeNet: An End-to-End System for Single Image Haze Removal."IEEE Transactions on Image Processing25.11(2016):5187-519.
Copyright © 2025 杨效禹, 张又升
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
