基于树莓派的智能车牌识别系统设计

Design of Intelligent License Plate Recognition System Based on Raspberry Pi

以立体车库为应用背景,为了提高车辆的智能化和网络化管理,设计一套基于树莓派的智能车牌识别系统。系统硬件部分以树莓派4B作为主控制端,以E18-D80NK红外传感器作为来车检测模块,以高清摄像头作为实时图像采集模块。车牌识别算法部分应用OpenCV库实现车辆图像预处理及初步的车牌定位分割操作;针对初步定位得到的易混淆非车牌区域,采用卷积神经网络进行过滤;最终用卷积神经网络实现车牌的字符识别并将识别结果显示在图形用户界面(Graphical User Interface,GUI)。实验结果表明,所设计的车牌识别系统准确率达到99%,识别效果良好。

Taking the stereo garage as the application background,a Raspberry Pi based intelligent license plate recognition system is designed to improve the intelligent and networked management of vehicles.The hardware part of the system uses Raspberry Pi 4B as the main console,E18-D80NK infrared sensor as the incoming vehicle detection module,and high definition camera as the real-time image acquisition module.The license plate recognition algorithm part applies OpenCV library to realize vehicle image preprocessing and preliminary license plate localization segmentation operation;for the preliminary localization of the easily confused non-license plate region,a convolutional neural network is used for filtering;finally,the convolutional neural network is used to realize the character recognition of the license plate and display the recognition results in the Graphical User Interface(GUI).Experimental results show that the license plate recognition system designed in this paper has an accuracy rate of 99%and good recognition effect.