摘要
本文主要针对高光谱图像的特点,利用波段间的几何信息高冗余性,通过小波分解去除高频的噪声和几何信息,保留低频的光谱信息。以其他波段的几何信息辅助噪声污染波段重构,经过相应的小波重构滤波器滤波,获得该波段图像的重建以进行消噪。
There is much correlative information among hyperspectral remote sensing image bands. We can use such character to denoise hyperspectral images. Based on wavelet transformation, wipe off noise band＇s high frequency signal which usually is noise and geometric information, and remain low frequency signal which is spectral information. Assisted by other band＇s high frequency geometric information, reconstruct noise band image with wavelet reconstruction filter to de-noise noise band image.
出处
《遥感信息》
CSCD
2005年第4期10-12,30,共4页
Remote Sensing Information
基金
国家"863"-2003AA131060项目支持
