摘要
深度学习在智能车行驶控制中起着关键作用。设计一种基于深度学习算法的智能车行驶控制系统,将车载摄像头和深度学习算法结合,利用U-Net网络模型和YOLO模型实现车道线检测和道路标识识别,同时设计有限状态机,完成智能车行驶控制。实验结果表明,本系统可以在模拟的道路场景中完成车道线和标识识别,实现车辆自主行驶控制,较传统计算机视觉方法在检测速度和检测效果上均得到提升。
Deep learning plays a key role in intelligent vehicle driving control system.The system is designed based on deep learning al-gorithm.The on-board camera and deep learning algorithm are combined to realize lane detection and road identification by using U-Net network model and YOLO model.At the same time,a finite state machine is designed to complete driving control.The experimental re-sults show that the system can realize lane line and sign recognition in the simulated road scene and realize the intelligent vehicle driv-ing control.Compared with the traditional computer vision method,the detection speed and detection effect of the designed sysem are improved.
作者
刘晨
叶浩
张德龙
赵兴哲
LIU Chen;YE Hao;ZHANG Delong;ZHAO Xingzhe(School of Computer Science and Engineering,Northeastern University,Shenyang Liaoning 110819,China)
出处
《电子器件》
CAS
北大核心
2023年第6期1645-1651,共7页
Chinese Journal of Electron Devices
基金
国家自然科学基金项目(62072085)。
关键词
深度学习
行驶控制
车道线检测
标识识别
deep learning
driving control
lane line detection
identification recognition
