基于垂直投影法的涉车案件车牌字符分割研究  

Research on License Plate Recognition in Vehicle Related Cases Based on Character Segmentation

作　　者：叶超 王震[1,1] YE Chao;WANG Zhen

机构地区：[1]中国刑事警察学院,辽宁沈阳110035

出　　处：《内蒙古公路与运输》2018年第1期48-53,共6页Highways & Transportation in Inner Mongolia

摘　　要：由交通事故引起的纠纷、停车场管理困难、车辆的丢失等社会性问题,以及随之引发的各类违法犯罪行为,给公安工作带来很多压力。促进智能交通系统日常化成为政府工作的重中之重,目前的车牌自动识别技术日趋完善,但算法的设计思路不尽相同,本文以车牌为研究对象,在分析车牌结构及型式特点的基础上,利用计算机图像处理技术,构建了车牌图像分割框架,建立了车牌图像灰度化、中值滤波、边框去除等预处理算法及基于垂直投影法的车牌图像分割模型,应用MATLAB集成了车牌图像分割系统。通过实际应用,本文建立的车牌分割具有较高质量的分割效果,且鲁棒性较强,为实现车牌自动识别提供了技术支撑。Social problems such as disputes caused by traffic accidents, difficulties in parking lot management and vehicle losses, and various illegal behaviors caused by it bring much pressure to public security work. Promoting the daily operation of the intelligent transportation system has become the most important part of the government＇s work. At present,the technologies of automatic recognition of license plate are becoming more and more perfect, but the design ideas of the algorithms are not the same. In this paper, taking license plate as the research object, based on the analysis of license plate structure and type characteristics, a license plate segmentation framework is constructed by using computer image processing technology, and the license plate image segmentation model is established based on gray level, median filter, edge removal and other preprocessing algorithms for license plate image and the vertical projection method. The license plate image segmentation system is integrated by MATLAB. Through practical application, the license plate segmentation established in this paper has high quality segmentation effect and strong robustness, which provides technical support for the realization of license plate recognition.

关 键 词：车牌识别系统 预处理 字符分割 数字图像处理技术 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]