基于树莓派的智能小车的设计与开发  

Design and development of intelligent cars based on the Raspberry Pi

作　　者：韩改宁 苏静池 张瑞斌 HAN Gaining;SU Jingchi;ZHANG Ruibin(School of Computer Science,Xianyang Normal University,Xianyang 712000,China)

机构地区：[1]咸阳师范学院计算机学院,陕西咸阳712000

出　　处：《电子设计工程》2024年第1期6-10,共5页Electronic Design Engineering

基　　金：国家自然基金项目(62073218);陕西省教育厅项目(17JK0826)。

摘　　要：针对疫情期间或者特殊行业无接触配送的需求,基于树莓派4B处理器及Raspberry操作系统,采用Python开发语言及TensorFlow、OpenCV和RPi.GPIO等实现开发。通过车载摄像头采集图像,利用OpenCV技术对图像进行处理,实现循迹功能;通过建立深度学习模型,对红绿灯数据集进行训练,利用训练好的模型实现红绿灯识别功能;通过红外避障模块实现道路避障功能。在带有红绿灯、障碍物和四个弯道的长方体模拟轨道进行测试,能准确地完成两圈行驶。该设计原理清晰,易于实现,在无接触配送、无人驾驶等方面具有很高的应用价值。For contactless distribution during the epidemic period or special industry requirements,based on the Raspberry Pi 4B processor and Raspberry operating system,using the Python development language and TensorFlow,OpenCV and RPi.GPIO and other libraries are used to implement the dev⁃elopment.The image is collected by the vehicle camera and processed by OpenCV technology to realize the tracking function.The traffic light dataset was trained by establishing a deep learning model,and the traffic light recognition function was realized by using the trained model.Through the infrared obstacle avoidance module to achieve the road obstacle avoidance function.The test was carried out on the cuboid simulation track with traffic light,obstacles and four curves,which could accurately complete two laps of driving.The design principle is clear,easy to implement,and has high application value in non⁃contact distribution,unmanned driving and so on.

关 键 词：智能小车 树莓派 TensorFlow RPi.GPIO 

分 类 号：TN4[电子电信—微电子学与固体电子学]