基于RBF神经网络的车牌自动识别系统设计与实现  

The Design and Apply of the Vehicle License Plate Recognition System Based on RBF Neural Network

作　　者：许智榜[1] 石晓瑛[1] 

机构地区：[1]华东交通大学电气与电子工程学院,江西南昌330013

出　　处：《南昌大学学报（工科版）》2009年第2期147-150,共4页Journal of Nanchang University(Engineering & Technology)

基　　金：国家自然科学基金资助项目(60533010)

摘　　要：车牌自动识别系统分为图像预处理、车牌定位、字符分割、字符识别4步。车牌准确定位是LPR系统中的关键,本文利用形态学变换对图像进行滤波聚类,HOUGH变换方法对车牌图像进行水平校正,基于RBF网络的方法识别字符。在DELPHI 7.0环境下设计开发的车牌自动识别系统,经检验取得满意的效果。The LPR （License Plate Recognition） system consists of four steps： image processing methods, license plate locating, character segmentation and character identification. License plate locating is an important step in a LPR system. This paper presents a method of locating and adjusting license plates, which combines the morphological method of image processing and the method of Hough transform. A character recognition algorithm based on RBF neural network is presented. The experiments proved that algorithm is fast and accurate. At last, a system of vehicle license identification system based on Delphi 7.0 was designed. The experimental results show the good performance of the LPR system.

关 键 词：车牌识别 RBF神经网络 HOUGH变换 字符分割 字符识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]