| 汽车牌照的提取方法 |
| |
| 引用本文： | 金玲玲,廖芹,汪刘一.汽车牌照的提取方法[J].华南理工大学学报(自然科学版),2002,30(7):95-98. |
| |
| 作者姓名： | 金玲玲 廖芹 汪刘一 |
| |
| 作者单位： | 1. 华南理工大学应用数学系,广东广州,510640
2. 华南农业大学工程学院,广东广州,510642 |
| |
| 摘 要： | 汽车牌照提取是汽车牌照自动识别的关键一步，本文中提出了两种提取方法，即利用扫描离差数据、有效谷峰点特征及先验知识来初步定位车牌区域；采用彩色分割及多级联合混合集成分类器的车牌提取技术，通过多层感知器网络对输入彩色图像进行彩色分割，将这两种方法联合使用对车牌定位准确率高、鲁棒性好，车牌定位的准确率达99.3%，具有很好的实用技术指标。
|
| 关 键 词： | 汽车牌照 提取方法 车牌定位 有效谷峰点 彩色分割 人工神经网络 混合集成分类器 车牌识别 字符提取 |
| 文章编号： | 1000-565X(2002)07-0095-04 |
| 修稿时间： | 2001年11月23 |
Algorithms to Detect Car Plate Position |
| |
| Jin Ling_ling,Liao Qin,Wang Liu_yi.Algorithms to Detect Car Plate Position[J].Journal of South China University of Technology(Natural Science Edition),2002,30(7):95-98. |
| |
| Authors: | Jin Ling_ling Liao Qin Wang Liu_yi |
| |
| Institution: | Jin Ling_ling 1 Liao Qin 1 Wang Liu_yi 2 |
| |
| Abstract: | Detecting car plate position is a key step in an automatic car plate recognition system. This paper proposes two methods for detecting car plate position from images: One is to utilize the difference data between scanning lines, property of effective peaking points and some pre_knowledge to basically determine the car plate area. Another method is to use color segmentation and multiplayer perceptron classifier. Multi_layer Percetron Networks are used for color image segmentation. Applying both methods together results in a high accurate position, and robustness. The accurate rate has reached as high as 99.3%, which is a good level for practical system. |
| |
| Keywords: | car plate position effective peak color segmentation artificial neural networks multiplayer perception classifier |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
