基于外观特征与神经网络的交通标志识别 |
| |
引用本文: | 朱正平,孙传庆,王秀丽,王阳萍.基于外观特征与神经网络的交通标志识别[J].自动化与仪器仪表,2009(1):60-63. |
| |
作者姓名: | 朱正平 孙传庆 王秀丽 王阳萍 |
| |
作者单位: | 1. 兰州城市学院计算机系,兰州,730070;兰州交通大学电子与信息工程学院,兰州,730070 2. 兰州城市学院计算机系,兰州,730070 3. 兰州交通大学电子与信息工程学院,兰州,730070 |
| |
基金项目: | 高等学样博士学科点专项科研基金 |
| |
摘 要: | 不同种类的交通标志具有特定的颜色及形状等外观特征,本文利用此特点设计了一个自动交通标志识别系统。该系统首先应用HIS彩色模型及标志的形状特征确定彩色图像中的标志区域及标志所属的种类。系统再应用自组织神经网络(S0MNN)进一步识别标志模式。实验证明了该方法的有效性与鲁棒性。
|
关 键 词: | 标志检测 标志识别 自组织神经网络 |
The traffic signs identification based on appearance charactristics and neural network |
| |
Abstract: | Different categories of traffic signs have different appearance characteristics of color and shape. Using the feature, an automatic traffic sign recognition system is desigened. Firstly, the system can appliy HIS color model and shape feature of traffic signs to determine sign regions in color images and the category of the signs. The self-organizing map neural network (SOMNN) is utilized to recognize an object in a determined category of objects. The experimental results show the feasibility and robustness of the proposed algorithm. |
| |
Keywords: | Sign detection Sign recognition Self-organizing map neural network |
本文献已被 维普 万方数据 等数据库收录! |
|