伪恒定物体(PIF)方法在遥感影像辐射匹配中的应用  

Application of Radiometric Scene Normalization Using Pseudoinvariant Features

作　　者：朱云燕[1] 贺彬[1] 赵碧云[1] 

机构地区：[1]云南省环境科学研究所,云南昆明650034

出　　处：《云南环境科学》2002年第2期12-14,共3页Yunnan Environmental Science

摘　　要：伪恒定物体 (PIF)方法基于影像上人造物体像元反射率的统计不变量 ,假定影像间同一人造物体像元的亮度值呈线性函数关系 ,通过数理统计方法 ,校正多时相、多波段影像之间的大气、光照和传感器参数差异 ,达成影像之间的辐射匹配。The technique of pseudoinvariant features corrects for atmospheric degradation,illumination effects,and sensor response differences in multitemporal multispectral images.The technique is based on the statistical invariance of the reflectance of man-made in-scene elements such as roads,rooftops and parking lots which covered with concrete or asphalt.Differences in the gray-level distributions of these invariant objects is assumed to be a linear function and is corrected statisically to perform the normalization.The technique has been used in the monitoring of Dianchi Lake.

关 键 词：伪恒定物体 PIF 遥感影像 辐射匹配 滇池 环境监测 

分 类 号：X87[环境科学与工程—环境工程] TP75[自动化与计算机技术—检测技术与自动化装置]