基于深度学习的道路识别与寻迹  

Road recognition and tracing based on deep learning

作　　者：李彦锡 周洋 李静峰 周琳习 陈昌达 范小兰[1] 赵春锋[1] Li Yanxi;Zhou Yang;Li Jingfeng;Zhou Linxi;Chen Changda;Fan Xiaolan;Zhao Chunfeng(Engineering Training Center,Shanghai University of Engineering Science,Shanghai 201620,China)

机构地区：[1]上海工程技术大学工程训练中心,上海201620

出　　处：《现代计算机》2023年第15期117-120,共4页Modern Computer

基　　金：上海工程技术大学创新创业项目(CS22240005):基于深度学习的道路识别与循迹。

摘　　要：根据第十七届全国大学生智能车竞赛百度完全模型组命题,对道路中经过精心设计的方案和多种算法进行研究,采用嵌入式树莓派系统,基于PaddlePaddle深度学习框架,运用Python语言进行图像处理,并利用卷积神经网络对管道上的物料进行分类,取得了令人满意的效果。According to the proposition of Baidu Complete Model Group in the 17th National College Student Smart Car Com-petition.The embedded raspberry pie system was chosen for the hardware configuration following the thorough examination of the overall scheme and its various algorithms.Utilizing the PaddlePaddle deep learning framework,the Python language was employed to complete image processing,while the convolution neural network design was utilized to classify the materials connected to the pipeline,resulting in satisfactory results.

关 键 词：形状识别 PaddlePaddle 卷积神经网络 PYTHON 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] U495[自动化与计算机技术—计算机科学与技术]