基于变化检测的国产高分系列卫星影像匀色研究  

An Image Dodging Algorithm for Multi-band Satellite Images Based on Change Detection

作　　者：张孝成[1,2] 胡勇 黄健[1,2] ZHANG Xiaocheng;HU Yong;HUANG Jian(Chongqing Institute of Surveying and Planning for Land Resources and Housing, Chongqing 400020,China;Chongqing Engineering Research Center of Land Use and Remote-Sense Monitoring, Chongqing 400020, China)

机构地区：[1]重庆市国土资源和房屋勘测规划院,重庆400020 [2]重庆市土地利用与遥感监测工程技术研究中心,重庆400020

出　　处：《测绘地理信息》2019年第4期102-106,共5页Journal of Geomatics

基　　金：住房和城乡建设部科技项目(2016-k8-054);重庆市国土房管科技项目(KJ-2015011)

摘　　要：遥感影像匀色处理是多源遥感影像拼接镶嵌前的重要处理步骤。基于遥感影像变化检测建立了一种多波段卫星影像的自动匀色方法:首先,对两景影像的重叠区域进行逐像元的光谱向量变化检测分析;然后,自动提取未变化的像元点;最后,以一景影像作为基准,利用未变化的像元点建立回归模型并对另一景影像进行匀色处理。高分一号、高分二号试验数据处理结果表明,该方法可有效消除或减小多源遥感影像之间的色彩差异。Image dodging is a crucial pre-processing step before image mosaicking. In this research, an automated image dodging procedure based on change detection is developed. First, a per-pixel spectral vector change detection is employed for overlapped areas of two satellite images, then the no-change pixels are extracted automatically. Third, based on those no-change pixels, a linear regression model is established through the unchanged pixel and meanwhile image dodging is conducted. The image dodging of GF-1 and GF-2 Satellite Images show that the image dodging procedure can effectively eliminate the radiation inconsistency between multi-temporal, multi-sensors images.

关 键 词：变化检测 匀色 高分一号 高分二号 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]