摘要
针对遥感蚀变信息提取中存在的不确定性和难点问题,文章基于光谱混合模型理论和随机变量概率分布理论,对多光谱遥感数据的空间几何结构特征进行了深入分析,探知了遥感图像多元数据集合的高维空间属于低维几何结构,并在此基础上,较系统地研究了背景、干扰与蚀变信息在特征空间中的分布关系,以及干扰地物类型的划分等问题,为遥感蚀变信息的准确提取提供了重要的技术思路。
In view of the uncertainty and difficulty of remote sensing alteration information extraction, based on spectrum mixed model theory and random variable probability distribution theory, the paper carries on thorough analysis to the muhi-spectrum remote sensing data＇s spatial geometry structural feature, finds out remote sensing image multi-data set higher dimensional space belong to lowers dimensional geometry structure. Based on this, systematically studies the distribution relations in the feature space of background, disturbance and alteration information, as well as disturbance ground object type division, provides the important technical thoughts for the remote sensing alteration information＇s accurate extraction.
出处
《地质与勘探》
CAS
CSCD
北大核心
2009年第1期46-51,共6页
Geology and Exploration
基金
“十一·五”科技支撑计划重大项目(编号:2006BAB01A06)
国土资源大调查项目(编号:1212010761503)资助
关键词
多(高)光谱数据空间结构
遥感蚀变信息
背景与干扰
二维散点图
multi（high） spectrum data ＇ s patial structure , remote sensing alteration information , background and disturbance , two dimensional scatter diagram
