基于锚点分类改进的自动驾驶车道线检测研究

doi:10.16097/j.cnki.1009-6744.2026.03.002

交通运输系统工程与信息 ›› 2026, Vol. 26 ›› Issue (3): 14-24.DOI: 10.16097/j.cnki.1009-6744.2026.03.002

基于锚点分类改进的自动驾驶车道线检测研究

黄凯^*a,b，谢子俊^a,b，李濠宇^a,b ，刘芯彤^a,b，刘志远^c

东南大学，a.仪器科学与工程学院；b.综合时空网络与装备技术全国重点实验室；c.交通学院，南京210096

收稿日期:2025-11-26 修回日期:2025-12-16 接受日期:2025-12-29 出版日期:2026-06-25 发布日期:2026-06-22
作者简介:黄凯（1990— ），男，江苏徐州人，副教授，博士。
基金资助:
国家自然科学基金青年科学基金(72301066)。

Improved Lane Line Detection in Autonomous Driving Based on Anchor Point Classification

HUANG Kai^*a,b, XIE Zijun^a,b, LI Haoyu^a,b, LIU Xintong^a,b , LIU Zhiyuan^c

a. School of Instrument Science and Engineering; b. State Key Laboratory of Comprehensive PNT Network and Equipment Technology; c. School of Transportation, Southeast University, Nanjing 210096, China

Received:2025-11-26 Revised:2025-12-16 Accepted:2025-12-29 Online:2026-06-25 Published:2026-06-22
Supported by:
Young Scientists Fund of the National Natural Science Foundation of China (72301066)。

摘要/Abstract

摘要： 近年来，车道线检测已成为自动驾驶车辆中的一项关键技术。然而，车道线检测受复杂环境因素影响较大，难以实现精确识别。为实时且高精度地在各种复杂环境下检测车道线，本文提出一种基于锚点分类的车道线检测模型（DG-UFLD）。首先，设计一种全局空间通道注意力模块（GSCA），该模块能够独特地捕捉多维全局信息，并增强特征表示，在捕捉多维全局信息和强化特征表示方面表现突出，进而提升模型整体性能；其次，在模型中引入动态蛇形卷积（DSConv）和三元组损失函数，进一步加强模型对细长目标的检测能力；最终，在模型整合中采用粗粒度车道网格分类替代细粒度分割，既能保持较高准确率，又能实现极高的检测速度。为验证本文提出的DG-UFLD算法的有效性，在两个公开车道线检测数据集（TuSimple和CULane）上进行实验。结果表明，相比于原始算法，所提出的DG-UFLD算法在保持极高检测速度的同时，在TuSimple数据集上的检测精度从原本的96.05%提升到96.45%；在CULane数据集上检测精度从原本的68.4%提升到75.3%。同时，与其他主流车道线检测网络相比，改进后的网络在极端情况下的车道线检测取得了良好的效果。验证结果证明本文方法能够快速和准确地检测车道线。

关键词: 智能交通, 车道线检测, 深度学习, 车辆前视图像, 自动驾驶, 计算机视觉

Abstract: In recent years, lane detection has become a key technology in autonomous vehicles. However, lane detection is difficult to achieve accurate recognition because it greatly affected by complex environmental factors. In order to detect the lane markings with high accuracy and effectiveness in various complex environments, this paper innovatively proposes a lane markings detection model DG-UFLD based on anchor point classification. Firstly, a novel Global Spatial Channel Attention Module (GSCA) is designed to improve the overall performance of the model, which can uniquely capture multidimensional global information and enhance feature representation. Secondly, dynamic snake convolution (DSConv) and triplet loss function are introduced into the model to further enhance its detection capability for slender targets. Finally, the integrated model adopts coarse-grained lane grid classification instead of fine-grained segmentation, which achieves extremely high detection speed with high accuracy. To verify the effectiveness of the proposed DG-UFLD algorithm, case studies are conducted on two publicly available lane detection datasets (TuSimple and CULane) in this paper. The results show that compared to the original algorithm, the proposed DG-UFLD algorithm improves the detection accuracy on the TuSimple dataset from 96.05% to 96.45% at extremely high detection speed; the detection accuracy on the CULane dataset increases from 68.4% to 75.3%. Meanwhile, compared with other mainstream lane detection networks, the improved network achieves better results in lane detection under extreme conditions. The verification results demonstrate that this method can quickly and accurately detect the lane markings.

Key words: intelligent transportation, lane detection, deep learning, vehicle front view image, autonomous driving, computer vision

中图分类号:

U495

黄凯, 谢子俊, 李濠宇, 刘芯彤, 刘志远. 基于锚点分类改进的自动驾驶车道线检测研究[J]. 交通运输系统工程与信息, 2026, 26(3): 14-24.

HUANG Kai, XIE Zijun, LI Haoyu , LIU Xintong , LIU Zhiyuan. Improved Lane Line Detection in Autonomous Driving Based on Anchor Point Classification[J]. Journal of Transportation Systems Engineering and Information Technology, 2026, 26(3): 14-24.

导出引用管理器 EndNote|Ris|BibTeX

链接本文: http://www.tseit.org.cn/CN/10.16097/j.cnki.1009-6744.2026.03.002

http://www.tseit.org.cn/CN/Y2026/V26/I3/14

参考文献

[1]刘登峰,郭文静,陈世海.基于内容引导注意力的车道线检测网络[J]. 浙江大学学报(工学版), 2025, 59(3): 451-459. [LIU D F, GUO W J, CHEN S H. Lane detection network based on content-guided attention[J]. Journal of Zhejiang University (Engineering Science), 2025, 59(3): 451-459.]

[2]GIRSHICK R. Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision, 2015: 1440-1448.

[3]HOU Y, MA Z, LIU C, et al. Learning lightweight lane detection CNNS by self attention distillation[C]// Proceedings of the IEEE/CVF International Conference on Computer Vision, 2019: 1013-1021.

[4]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.

[5] NEVEN D, DE BRABANDERE B, GEORGOULIS S, et al. Towards end-to-end lane detection: An instance segmentation approach[C]//2018 IEEE Intelligent Vehicles Symposium (IV), IEEE, 2018: 286-291.

[6]LONG J, SHELHAMER E, DARRELL T. Fully convolutional networks for semantic segmentation[C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2015: 3431-3440.

[7]PAN X, SHI J, LUO P, et al. Spatial as deep: Spatial CNN for traffic scene understanding[C]. Proceedings of the AAAI Conference on Artificial Intelligence, 2018.

[8]ZHAO H, SHI J, QI X, et al. Pyramid scene parsing network[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 2881 2890.

[9] 田会娟,刘嘉伟,翟佳豪,等.基于多入侵线的视频车速检测方法[J]. 交通运输系统工程与信息,2022,22 (1): 49-56, 84. [TIAN H J, LIU J W, ZHAI J H, et al. Video vehicle speed detection method based on multi intrusion lines[J]. Journal of Transportation Systems Engineering and Information Technology, 2022, 22(1): 49-56, 84.]

[10] QIN Z, WANG H, LI X. Ultra fast structure-aware deep lane detection[C]//European Conference on Computer Vision, Cham: Springer International Publishing, 2020: 276-291.

[11] 辛琪, 齐嘉乐,荚胜琪,等.基于多特征逐尺度回归的车道线检测模型[J].中国公路学报,2025,38(6): 313 323. [XIN Q, QI J L, JIA S Q, et al. Lane detection model based on multi-feature scale-wise regression[J]. China Journal of Highway and Transport, 2025, 38 (6): 313 323.]

[12] QI Y, HE Y, QI X, et al. Dynamic snake convolution based on topological geometric constraints for tubular structure segmentation[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision, 2023: 6070-6079.

[13] WANG Y, JING Z, JI Z, et al. Lane detection based on two-stage noise features filtering and clustering[J]. IEEE Sensors Journal, 2022, 22(15): 15526-15536.

[14] 杨达, 刘家威,郑斌,等.基于时域卷积网络与注意力机制的车辆换道轨迹预测模型[J].交通运输系统工程与信息,2024, 24(2): 114-126. [YANG D, LIU J W, ZHENG B, et al. Vehicle lane change trajectory prediction model based on temporal convolutional network and attention mechanism[J]. Journal of Transportation Systems Engineering and Information Technology, 2024, 24 (2): 114-126.]

[15] CHEN W, CHEN X, ZHANG J, et al. Beyond triplet loss: A deep quadruplet network for person re-identification [C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2017: 403-412.

基于锚点分类改进的自动驾驶车道线检测研究

Improved Lane Line Detection in Autonomous Driving Based on Anchor Point Classification

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

[1]	郭力铭, 刘硕, 郑建风. 复合航道调度双目标优化模型与改进深度学习算法[J]. 交通运输系统工程与信息, 2026, 26(3): 72-82.
[2]	王正礼, 唐子墨, 郑振杰. 融合检索增强生成与思维链的交通事故责任认定大模型构建[J]. 交通运输系统工程与信息, 2026, 26(3): 83-92.
[3]	姜晓红, 仲韫豪, 肖靖沂, 邢吉平, 李家伟, 华晶雯. 考虑即时零售订单的城乡公交响应式客货联运调度[J]. 交通运输系统工程与信息, 2026, 26(3): 156-165.
[4]	蒋贤才, 魏贺迪, 张馨月. 基于双智能体的交通信号与车辆轨迹动态权重联合优化[J]. 交通运输系统工程与信息, 2026, 26(3): 192-202.
[5]	徐晓美, 符蒙, 赵峻伟, 张涌. 混合交通流场景下基于改进人工势场法的队列协同换道控制[J]. 交通运输系统工程与信息, 2026, 26(3): 214-225.
[6]	张建华, 公佳豪, 张文会. 智能网联汽车复用公交车道协同控制研究[J]. 交通运输系统工程与信息, 2026, 26(3): 226-234.
[7]	曹倩霞, 陈世文, 吕松涛, 王大为. 复杂交通场景基于边缘特征增强的长队列排队检测[J]. 交通运输系统工程与信息, 2026, 26(3): 235-246.
[8]	郑展骥, 廖方正, 李燊, 冯昌奎, 凃强, 张河山, 徐进. 无人机夜间航拍视角下小目标车辆精确检测方法[J]. 交通运输系统工程与信息, 2026, 26(3): 247-258.
[9]	田君豪, 邢璐, 廖世豪, 桂瑰 , 蒋小晴. 基于地理自编码与跨域迁移的公交出行需求分层聚类方法[J]. 交通运输系统工程与信息, 2026, 26(3): 302-314.
[10]	刘少博, 苏蔚. 基于强化学习的地铁站客流动态管控策略研究[J]. 交通运输系统工程与信息, 2026, 26(3): 338-347.
[11]	王亦兵, 陈安妮, PAPAGEORGIOU Markos, 余宏鑫, 郭静秋, 章立辉. 智能网联无车道线城市道路内边界与交通信号协同控制[J]. 交通运输系统工程与信息, 2026, 26(2): 81-90.
[12]	谷远利, 宇泓儒, 陈龙, 邓社军, 陆文琦. 网联自动驾驶车辆专用车道动态宏微观协同部署方法[J]. 交通运输系统工程与信息, 2026, 26(2): 125-136.
[13]	孙健, 纪裕伟, 于珂伟, 李子豪, 赵昱霖. 融合双注意力机制的快速路协同深度强化学习方法[J]. 交通运输系统工程与信息, 2026, 26(2): 137-147.
[14]	万平, 陈盈, 邓鑫焰, 马晓凤. 基于高斯混合隐马尔可夫模型的路怒攻击性驾驶行为辨识[J]. 交通运输系统工程与信息, 2026, 26(2): 268-279.
[15]	刘涛, 李林. 融合贝叶斯优化与深度学习的机场公交专线短时客流预测[J]. 交通运输系统工程与信息, 2026, 26(2): 300-308.