汽车工程 ›› 2025, Vol. 47 ›› Issue (10): 1905-1913.doi: 10.19562/j.chinasae.qcgc.2025.10.006

• • 上一篇    

基于双分支实例分割网络的复杂环境车道线检测方法

王平1(),罗哲1,查云飞2,张义1,唐友名3   

  1. 1.厦门理工学院机械与汽车工程学院,厦门 361024
    2.福建理工大学机械与汽车工程学院,福州 350118
    3.浙江科技大学机械与能源工程学院,杭州 310023
  • 收稿日期:2024-11-26 修回日期:2025-04-24 出版日期:2025-10-25 发布日期:2025-10-20
  • 通讯作者: 王平 E-mail:pipiwang3487@xmut.edu.cn
  • 基金资助:
    国家自然科学基金(51705441);国家重点研发计划子课题(2023YFB3406500);福建省中青年教师教育科研项目(JAT210349)

Lane Detection for Complex Environment Based on Two-Branch Instance Segmentation Networks

Ping Wang1(),Zhe Luo1,Yunfei Zha2,Yi Zhang1,Youming Tang3   

  1. 1.School of Mechanical and Automotive Engineering,Xiamen University of Technology,Xiamen 361024
    2.School of Mechanical and Automotive Engineering,Fujian University of Technology,Fuzhou 350118
    3.School of Mechnical & Energy Engineering,Zhejiang University of Science & Technology,Hangzhou 310023
  • Received:2024-11-26 Revised:2025-04-24 Online:2025-10-25 Published:2025-10-20
  • Contact: Ping Wang E-mail:pipiwang3487@xmut.edu.cn

摘要:

针对自动驾驶汽车在车道线数量识别缺失、复杂环境下车道线分割精度不足的问题,提出一种双分支实例分割网络结构的车道线检测方法。首先,该方法使用编码-解码框架提高细节恢复能力,支持多层次特征融合;其次,融合特征金字塔网络与高级残差网络,提升模型对上下文信息和深层次语义的理解,有效提取复杂车道线中的语义特征;然后,引入SE模块和加权最小二乘拟合模块,加强模型整体特征表达能力和泛化能力,提高模型的灵活度与准确率,并在不损失模型计算性能的前提下,增强车道线的几何形态预测;最后,该算法在CULane和TuSimple数据集进行实验测试的F1分别达到76.0%和96.9%。实验结果表明,该方法在光照变化、遮挡和道路损伤、多车道等复杂环境中可获得良好的检测性能,有效提高车道线检测精度。

关键词: 实例分割, 车道线检测, 特征金字塔网络, 自动驾驶

Abstract:

For the problem of lack of lane line quantity recognition and insufficient accuracy of lane line segmentation in complex environment for self-driving cars, a lane line detection method with a two-branch instance segmentation network structure is proposed. Firstly, the method uses an encoding-decoding framework to improve the detail recovery ability and support multilevel feature fusion. Secondly, the fusion of feature pyramid network and advanced residual network improves the model's understanding of contextual information and deep semantics, which efficiently extracts the semantic features in complex lane lines. Then, the introduction of the SE module and the weighted least-squares fitting module strengthens the model's overall feature expression and generalization ability so as to improve the flexibility and accuracy of the model, and enhance the geometric shape prediction of lane lines without losing the computational performance of the model. Finally, the F1 of the algorithm experimentally tested on CULane and TuSimple datasets reaches 76.0% and 96.9%, respectively, and the experimental results show that the method can obtain good detection performance in complex environment such as light change, occlusion and road damage, and multiple lanes, and effectively improves the lane line detection accuracy.

Key words: instance segmentation, lane detection, feature pyramid network, autonomous driving