基于深度学习的隧道车道识别优化方法

Tunnel Lane Recognition Optimization Method Based on Deep Learning

文章提出一种基于深度学习的隧道车道识别优化方法,并构建自制数据集进行了实验验证。首先,研究一个基于深度学习的隧道车道识别框架;其次,针对框架的不足,提出预处理数据以增强数据可学习性,引入MobileNetV3模型有效捕捉隧道内复杂车道结构;最后,通过自制数据集进行实验验证。实验结果显示,在隧道环境中,该方法能够准确识别车道,具有良好的表现。

This paper proposes an optimization method for tunnel lane recognition based on deep learning,and constructs a self-made data set for experimental verification.First,a tunnel lane recognition framework based on deep learning was studied.Secondly,in view of the shortcomings of the framework,data preprocessing was proposed to enhance data learnability,and the MobileNetV3 model was introduced to effectively capture the complex lane structure in the tunnel;Finally,experimental verification was carried out through a self-made data set.Experimental results show that in a tunnel environment,this method can accurately identify lanes and has good performance.