摘要
提出了一种对高光谱图像的真实性进行检验的方法,借助地物光谱仪获取的光谱分辨率更高的同步地物光谱数据,使用光谱角度匹配的方法对高光谱图像光谱信息定量化分析,以此结果作为高光谱图像数据的真实性判据。计算前对参考光谱重采样及对光谱向量进行一阶微分变换,能够减少噪声影响、提高分析结果的可信度。实际飞行数据实验结果证明了这种方法的可行性。
A new hyperspectral image validation method was proposed. In virtue of higher-resolution hyperspectal data collected by spectrometer, quantitative analysis of the validation of the spectral information was operated on the basis of hyperspectral image by using the method named spectral angle mapper. Then the result of the analysis was referred as criterion of the verification of hyperspectral image. To improve the reliability of the reference spectral vector, which was resampled before being calculated and 1-order differential was employed to reduce noise from spectral vectors. The experiment result validated the feasibility of the method.
出处
《激光与红外》
CAS
CSCD
北大核心
2007年第1期90-93,共4页
Laser & Infrared
基金
国家863计划资助项目(863-13-2002AA130010-1-4)
关键词
遥感
高光谱图像
真实性检验
向量相关
光谱角度匹配
remote sensing
hyperspectral image
validation
vectors correlation
spectral angle mapper
