摘要 在遥感高光谱图像获取过程中,由于地面大气之间的辐射传输过程非常复杂,导致高光谱图像降质严重,需要进行辐射校正。然而,如何评价辐射校正方法的性能是一个很重要的问题。本文研究所开发的辐射校正软件-HRC 与商业软件Envi 的精度对比以及基于地面光谱采样数据的HRC 辐射校正性能分析。19281
本文首先给出了辐射校正性能综合评价实施方案的设计与处理流程; 设计了一种基于kappa 一致性指数分析的辐射校正性能评价的方法。方法步骤包括对校正图像进行监督分类,计算样本总样,建立误差矩阵和计算kappa 一致性指数等步骤。 本文开发了相应的辐射校正性能评价程序,通过与商业软件 Envi 和地面光谱数据的一致性分析,仿真实验表明:辐射校正软件-HRC 与商业软件Envi 具有较好的一致性,且校正后的数据与地面采集的光谱数据的一致性概率高。
关键词 高光谱图像; 辐射校正; 误差矩阵; kappa 一致性指数
Title Research on hyperspectral images' radiation correction performance evaluation methods based on the measured spectra samples of ground's typical uniform surface features
Abstract
During the process of remote hyperspectral images' obtaining,because of
the complexity of ground-atmosphere radiation transfer process,it causes
hyperspectral images's quality decreases a lot.So we need to go on
radiation correction.However,how to evaluate the performance of radiation
correction methods is a very important question.The paper studies on the
accuracy comparison between the developed radiation correction
software--HRC and the business software--Envi.This paper also makes a
study on evaluation of HRC radiation correction performance based on
spectral data sampled on the ground.
At first,this paper introduces the design and process flow of radiation
correction's evaluation implementation.In addition ,this paper also
design a kind of method of radiation correction performance evaluation
based on kappa consistency index analysis. The method includes supervised
classification for correcting image, calculating the total sample
samples ,establishing consistency error matrix and calculating kappa index
and etc.
This paper develops a corresponding program of radiation correction
performance's evaluation.With the kappa analysis of Envi and ground
spectral data,the simulate experiment shows,the consistency index between HRC and Envi is good.So is the index between corrected data and ground
sampled spectral data.
Keywords hyperspectral image; radiation correction; error matrix;
kappa consistency index
目 次