J4 ›› 2006, Vol. 28 ›› Issue (8): 50-52.
• 论文 • 上一篇 下一篇
陈修桥[1] 胡以华[2] 蔡晓春[2] 张军[2]
出版日期:
发布日期:
Online:
Published:
摘要：
针对自适应神经模糊推理系（ANFIS）的结构特点，本文应用多个ANFIS组合的网络来解决多目标分选问题。为了给ANFIS赋予一个合适的初始状态，减法聚类被用于预处理输 入特征数据，实现了网络结构和初始参数的优化。基于ANFIS的红外航母和大型舰船目标分选仿真实验证明，其学习速度快、目标分选准确。
关键词: ANFIS 聚类 识别 红外目标
Abstract:
As for the characteristics of the adaptive network-based fuzzy inference system （ANFIS）, the paper applies a network based on multiple ANFISs to solve the problem of multiple target classification. The input feature data is processed first by subtractive clustering, which can optimize the network st ructure and initialize the parameters of ANFIS. The simulation of infrared remote sensing target classification based on ANFIS and subtractive clustering proves that the algorithm can efficiently and accurately classify targets.
Key words: ANFIS, cluster, classification, intrared target
陈修桥[1] 胡以华[2] 蔡晓春[2] 张军[2]. 基于ANFIS和减法聚类的遥感红外目标分选[J]. J4, 2006, 28(8): 50-52.
0 / / 推荐
导出引用管理器 EndNote|Ris|BibTeX
链接本文: http://joces.nudt.edu.cn/CN/
http://joces.nudt.edu.cn/CN/Y2006/V28/I8/50
