NOAA卫星图像云点检测的三维统计图集合判别  

A METHOD FOR DETECTING CLOUD PIXELS ON NOAA-AVHRR IMAGE USING THREE-DIMENSION STATISTIC GRAPH AND SET DISCRIMINATION

作　　者：杨俊泉[1] 沈建中[1] 莫伟华[2] 张远飞[3] 陈尚文[1] 林少雄[4] 

机构地区：[1]广西大学林学院 [2]广西气象台 [3]中国有色总公司矿产地质研究院 [4]广西气象局

出　　处：《国土资源遥感》1997年第4期54-59,共6页Remote Sensing for Land & Resources

摘　　要：在应用ＮＯＡＡ－ＡＶＨＲＲ图像对地球资源环境进行监测的研究中，云点检测是一项基本工作。本文以实例分析云的图像特征，选取特征指标建立像元分布三维统计图，直接在统计图上观察云点集群分布阈值，并以此作为云点集合判别的条件，以实现自动检测。此法灵敏度高，稳定性好，且直观方便。The detection of cloud pixels is the essential for monitoring resources and environment of the earth using NOAA-AVHRR image. The paper analyses the feature of the clouds on NOAA-AVHRR data by instances, draws 3D-statistic graph according to the feature indexes, and decides the thresholds of cloud pixels directly from the graph. The thresholds are regarded as the setting condition for detecting the cloud pixels automatically. The advantages of above method are higher sensitivity and better stability and that the viewing for the distribution of cloud pixels is more directly.

关 键 词：NOAA-AVHRR图像 云检测 三维统计图 集合判别 

分 类 号：P412.15[天文地球—大气科学及气象学] P455