CAJ | 학술논문

在车辆牌照自动识别系统中，因自然因素或采样因素使得原本规则的印刷体字符产生畸变，给字符识别带来了很大困难，利用 BP 神经网络对车牌进行识别得到广泛的应用，如何提高车牌识别的准确率和速度是车牌识别系统最根本的问题。本文用动量因子和自适应学习速率对传统 BP 网络进行改进。该算法简便，识别率高，可适用于多种高噪声环境中的字符识别。
재차량패조자동식별계통중，인자연인소혹채양인소사득원본규칙적인쇄체자부산생기변，급자부식별대래료흔대곤난，이용 BP 신경망락대차패진행식별득도엄범적응용，여하제고차패식별적준학솔화속도시차패식별계통최근본적문제。본문용동량인자화자괄응학습속솔대전통 BP 망락진행개진。해산법간편，식별솔고，가괄용우다충고조성배경중적자부식별。
In the vehicle license plate automatic recognition system,it is difficult to identify the charac-ters due to natural or sampling factors,which makes the original printed characters be distorted. BP neural network has been widely used to identify the license plate.How to improve the accuracy and speed of the license plate recognition is the most fundamental problem in license plate recogni-tion system.In this paper,we use momentum factor and adaptive learning rate to improve the tra-ditional BP network.Being simple,with high recognition rate,the algorithm can be applied to character recognition in a variety of high-noise environments.