摘要
针对超谱遥感图像的高数据维给图像进一步处理带来的困难,本文构造了波段选择方法的数学模型,该方 法基于统计学原理,通过选择信息量大并且与其它波段相关性小的波段来降低超谱数据维数。本文将降低后的超谱数据进 行小波融合与K．均值非监督分类。分类结果表明,该波段选择的方法能够将保留信息丰富的波段,分类效果与使用原始 波段相比有所提高,计算复杂度大大降低。
Aim at the problems brought by high dimensions of hyperspectral remote sensing image, the algorithm model of band selection that based on statistics was constructed in this paper, which reduces dimensions by selecting high informative and low correlative band. Then wavelet fusion and K-mean unsupervised classification were carded on the dimensionally reduced image. The results of the classification showed the new approach can contain high informative bands and the effect of classification is better than raw image, computing complication is also decreased rapidly.
出处
《信号处理》
CSCD
北大核心
2005年第6期676-680,共5页
Journal of Signal Processing
基金
本课题受哈尔滨市学科后备带头人基金哈尔滨工程大学基础科研基金资助
关键词
超谱遥感图像
波段选择
小波融合
分类
hyperspectral remote sensing image
band selection
wavelet fusion
classification
