首页 | 本学科首页   官方微博 | 高级检索  
     检索      

基于Tabu搜索的高光谱影像特征选择
引用本文:杨哲海,张雅争,宫大鹏,李之歆,韩建峰.基于Tabu搜索的高光谱影像特征选择[J].海洋测绘,2006,26(4):11-14.
作者姓名:杨哲海  张雅争  宫大鹏  李之歆  韩建峰
作者单位:1. 解放军信息工程大学,测绘学院,河南,郑州,450052;65015部队,辽宁,大连,116023
2. 海军大连舰艇学院,辽宁,大连,116018
3. 65015部队,辽宁,大连,116023
基金项目:高比容电子铝箔的研究开发与应用项目
摘    要:在高光谱影像的分类过程中,如何有效地降低特征空间的维数,又能保证原始数据所包含的丰富地物信息,是一项十分重要而繁琐的工作。在分析了传统降维方法所面临问题的基础上,将禁忌(Tabu)搜索算法引入到高光谱影像的特征选择研究,指出由于Tabu搜索算法所具有的良好全局寻优能力,因而在该类影像的降维研究中有着广阔的应用前景。实验表明,将Tabu搜索算法获取的波段进行高光谱影像分类,在求解的时间和分类结果的精度上都可达到令人满意的效果。

关 键 词:高光谱  遥感  特征选择  禁忌搜索
文章编号:1671-3044(2006)04-0011-04
收稿时间:2006-01-12
修稿时间:2006-03-28

Feature Selection in Hyperspectral Classification Based on Tabu Search Algorithm
YANG Zhe-hai,ZHANG Ya-zheng,GONG Da-peng,LI Zhi-xin,HAN Jian-feng.Feature Selection in Hyperspectral Classification Based on Tabu Search Algorithm[J].Hydrographic Surveying and Charting,2006,26(4):11-14.
Authors:YANG Zhe-hai  ZHANG Ya-zheng  GONG Da-peng  LI Zhi-xin  HAN Jian-feng
Institution:1. Institute of Surveying and Mapping, Information Engineering University, Zhengzhou, Henan,450052; 2. Dalian Naval Academy, Dalian, Liaoning, 116018; 3. 65015 Troops, Dalian, Liaoning, 116023
Abstract:In hyperspectral classification,it is necessary to reduce both the number of input dimensionality and maintain the information contained in original hyperspectral data.Based on analyzing the limitations of the traditional dimensionality reduction methods,the paper introduced the Tabu Search into hyperspectral feature selection.Also,the author pointed out that due to the advantage of solving the global optimal problems,Tabu Search algorithm will be widely used in hyperspectral feature reduction.Experimental results demonstrate that Tabu Search is very helpful to achieve good classification performance by selecting the most useful bands which will be used in classifiers.
Keywords:hyperspectral  remote sensing  feature selection  Tabu search
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号