摘要
图像识别的应用场景日益复杂,尤其是在摄像设备成像不佳的情况下,对高速行驶或横向移动的车辆进行拍照登记车牌的应用需求已非常广泛,比传统静止状态下车牌识别其识别难度更大。因此,设计一种在复杂环境下的车牌识别系统监控意义重大。本文运用卷积神经网络构建网络模型,通过汽车牌照照片去噪、二值化等预处理、车牌字符切割,按照车牌省份、城市代号、车牌编号三部分分别训练网络模型。经测试,该模型实现汽车牌照自动识别,具有准确率高、速度快和成本低等优点,具有良好的应用前景。
The application scene of image recognition is becoming more and more complex,especially in the case of poor imaging of camera equipment.The application demand of photographing and registering license plates for high-speed or horizontally moving vehicles has been very extensive,which is more difficult than traditional license plate recognition in static state.It is of great significance to design a license plate recognition system in complex environment.In this paper,convolutional neural network is used to build the network model,and the network model is trained according to the three parts of license plate province,city code and license plate number through pretreatment such as car license plate photo denoising,binarization and license plate character cutting.After testing,the model realizes the automatic recognition of vehicle license plate,which has the advantages of high accuracy,high speed and low cost,and has a good application prospect.
作者
徐金荣
郭彩萍
Xu Jinrong;Guo Caiping(Department of Electronic Engineering,Taiyuan Institute of Technology,Taiyuan Shanxi 030008,China)
出处
《山西电子技术》
2023年第1期50-52,55,共4页
Shanxi Electronic Technology
基金
山西省高校科技创新项目(2019L0924)
太原工业学院教研项目(2019YJ21452)。