基于深度学习的无人驾驶汽车车道线检测方法  

Lane Line Detection Method of Driverless Car Based on Deep Learning

作　　者：高扬[1] 王晨 李昭健 GAO Yang;WANG Chen;LI Zhao-jian(Automobile College, Chang’an University, Xi’an 710000, China;Automobile College, Xi’an 710000, China;Key Laboratory for Automotive Transportation Safety Enhancement Technology of the Ministry of Communication, Xi’an 710000)

机构地区：[1]长安大学汽车学院,西安710000 [2]汽车运输安全保障技术交通行业重点实验室,西安710000

出　　处：《科学技术与工程》2021年第24期10401-10406,共6页Science Technology and Engineering

基　　金：国家自然科学基金(61503043);陕西省自然科学基金(2019JLP-07,1019JM-309);中央高校基金(3100102229103)。

摘　　要：车道线检测是实现当前汽车辅助驾驶和未来无人驾驶汽车的关键,深度学习技术在近年来迅猛发展,在图像识别、图像分割、语音识别及数据预测等方面都取得了出色成绩。结合深度学习技术对无人驾驶汽车环境感知中的车道线检测进行了相应的研究,提出一种基于深度学习的车道线识别算法。对比研究已有算法,针对其中的信息融合问题,提出了一种新的特征图上下文信息融合方法,将该方法与VGG(Visual Geometry Group)网络相结合提出融合上下文信息的车道线识别网络VGG-FF,进一步加入空洞卷积提出融合空洞卷积及上下文信息的车道线识别网络VGG-FFD。将该网络模型在公开数据集以及自制数据集上进行了性能测试,实验结果表明该模型具有良好的识别效果。Lane line detection is the key to the realization of current car-assisted driving and future driverless cars.Deep learning technology has developed rapidly in recent years and has achieved excellent results in image recognition,image segmentation,speech recognition,and data prediction.Deep learning technology is combined and corresponding research on lane line detection in driverless car environment perception is carried out,a lane line recognition algorithm based on deep learning was proposed.Comparative study of existing algorithms,given the information fusion problem,a novel feature map context information fusion method was proposed.The VGG(visual geometry group)network was combined with the method,lane line recognition network VGG-FF fused with contextual information was proposed,and the dilated convolution was further added,lane line recognition network VGG-FFD combining hole convolution and contextual information was proposed.The network model was compared with the public data set and self-made data set.The experimental results show that the model has a good recognition effect.

关 键 词：无人驾驶技术 车道线检测 深度学习 信息融合 VGG网络 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]