基于左上边缘点检测的快速车牌定位算法  

Fast Algorithm for Vehicle License Plate Location Based on Left- up Edge- point Detection

作　　者：王剑书[1] 樊养余[1] 张辰锐[1] 

机构地区：[1]西北工业大学电子信息学院,陕西西安710129

出　　处：《计算机仿真》2013年第11期143-147,共5页Computer Simulation

摘　　要：研究车牌定位优化问题。实际道路交通环境下采集的车辆图片存在复杂的干扰,造成现有车牌定位算法定位率不高和运行速度慢的问题。为减少干扰的影响,提高定位精度并减少运行时间,提出了基于左上边缘点检测的快速车牌定位算法。方法提出了左上边缘点检测技术,缩小了车牌的全局搜索范围,抑制了干扰;构造的车牌评价函数,利用了车牌区域的多重特征,能更准确地表达车牌区域。对实际道路交通中采集的车辆图像进行实验,获得了较高的定位率与较少的运行时间,证明该算法具有较高的精度、良好的实时性和强鲁棒性。The algorithm for vehicle license plate location was studied. Images captured in practical road traffic u- sually contain complex interference, which might cause low precision and slow speed of current license plate location algorithm. In order to reduce the influence of interference and improve the accuracy and reduce the running time, a fast algorithm for vehicle license plate location based on left-up edge-point detection was proposed. In this method, we used the technique of left-up edge-point detection, so that the scope of global search was narrowed and the inter- ference was reduced. We constructed license plate evaluation function using the multiple features of the license plate area, so that the license plate area was described more accurately. From the experiment made on images captured in practical road traffic, higher positioning rate and less running time were obtained, which proves that the algorithm is of high location precision ratio, good real-time and strong robustness.

关 键 词：车牌定位 左上边缘点 边缘检测 车牌评价函数 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]