【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。更多还原

【Abstract】 According to the proposition of Baidu Complete Model Group in the 17th National College Student Smart Car Competition. 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.更多还原