| 基于DIGNET网络的数据融合方法 |
| |
| 引用本文： | 牛丽红,苏秉华,倪国强.基于DIGNET网络的数据融合方法[J].光学技术,2006,32(2):171-173. |
| |
| 作者姓名： | 牛丽红 苏秉华 倪国强 |
| |
| 作者单位： | 1. 深圳大学,光电子学研究所,广东,深圳,518060
2. 北京理工大学,信息科学技术学院光电工程系,北京,100081 |
| |
| 摘 要： | 针对数据融合和目标识别的特点,提出了基于DIGNET自组织聚类人工神经网络的数据融合方法。考虑到多传感器系统测量多个参量的特点,用并行的子网络结构代替中间隐层,实现了基于决策层的信息融合目标识别。利用仿真数据对基于DIGNET的数据融合方法进行了实验研究。实验结果表明,该方法具有数据正确分类率高和抗噪能力强等优点,有效地实现了融合识别。将该方法应用于前视红外和可见光双传感器目标跟踪系统的数据融合识别是可行的。
|
| 关 键 词： | 数据融合 人工神经网络 目标识别 非监督学习算法 |
| 文章编号： | 1002-1582(2006)02-0171-03 |
| 收稿时间： | 2005/7/27 |
| 修稿时间： | 2005年7月27日 |
A novel DIGNET network based data fusion approach |
| |
| NIU Li-hong,SU Bing-hua,NI Guo-qiang.A novel DIGNET network based data fusion approach[J].Optical Technique,2006,32(2):171-173. |
| |
| Authors: | NIU Li-hong SU Bing-hua NI Guo-qiang |
| |
| Abstract: | In view of the characteristics of data fusion and target recognition,a novel self-organizing clustering artificial neural network(DIGNET) based data fusion approach is proposed.Taking into consideration of multi-parameters of multi-sensor system,data fusion at decision level for target recognition was realized with parallel sub-networks.The fusion scheme based on DIGNET was studied using simulated data.The experimental results show that high correct classification rate and good tolerance to noise interference can be achieved for fusion recognition.The fusion architecture is used for a target tracing system of FLIR,and TV camera and the results indicate that it is workable. |
| |
| Keywords: | data fusion artificial neural networks target recognition unsupervised-learning algorithm |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
