摘要
PROBA是欧空局于2001年发射的一颗最小的对地观测卫星,CHRIS是搭载于PROBA平台上的最主要的成像光谱分光计,具有五个成像模式,以其卓越的光谱空间分辨率及多角度的优势为着不同的研究目的分别对陆地、海洋及内陆水体进行成像。但CHRIS/PROBA数据本身被严重的条带噪声所污染,尤其是垂直条带噪声,因而去除条带噪声则成为使用CHRIS数据的首要与处理工作。本文描述了几种去除条带方法,包括传统的矩匹配方法,新的迭代辐射纠正方法,校正因子法及比值法。由于这些方法基于不同的理念,各个方法的去噪效果必定有所差异,本文目的是通过比较选出其中最好的一种方法。每种方法的结果从两方面进行了比较、评定:定性的视觉效果及图像变换前后的定量的灰度偏移值,比较结果列成表格,从而挑选出最优秀的一种方法。
PROBA is the ESA＇s smallest earth observation satellite launched in 2001, and CHRIS is the main imaging instrument loaded on the PROBA platform, which have five image modes and can image the land, ocean and inland water bodies with its outstanding hyperspectral and high spatial resolution and multiple angular image ability for different purposes. But CHRIS/PROBA data is contaminated seriously by strip noise, especially the vertical strips, so destriping is the primary step of work before using it. This paper describes several destriping methods including conventional global, linear detector matching method, new iterative radiometric correction method, correction factor method and ratio reverse method and the goal of the paper is to select the best one. For these methods are designed based on different ideas, it is sure that their effects of de-striping are not the same between each other. Finally, the effect of every method is compared in two ways： visual effect and quantitative disparity before and after transformation to the original image, and the quantitative variance for different thresholds is tabled , then the best method is selected out.
出处
《遥感信息》
CSCD
2006年第6期36-39,49,共5页
Remote Sensing Information
基金
科技部科研院所社会公益研究专项(2003DIA6N015及2004DIB3J107)
