基于动态阈值法的NOAA系列卫星云检测结果检验  

Validation of NOAA/AVHRR Cloud Detections by an Automated Dynamic Threshold Cloud-Masking Algorithm

作　　者：武艳[1] 银燕[1] 师春香[2] 何立[3] 胡婷婷[1] 孟俊耀 

机构地区：[1]南京信息工程大学大气物理学院,江苏南京210044 [2]国家卫星气象中心,北京100081 [3]广西区气象减灾研究所,广西南宁530022 [4]东华软件股份公司,北京100081

出　　处：《高原气象》2012年第3期745-751,共7页Plateau Meteorology

基　　金：财政部/科技部公益类行业专项(GYHY200706005;GYHY200906036;GYHY200706036);中国气象局新技术推广项目(CMATG2008Z04);国家科技支撑计划课题(2006BAC12B07)共同资助

摘　　要：采用我国地面报资料、国际船舶报资料和云卫星(CloudSat)产品,分别对多通道动态阈值云检测方法的云检测结果进行了检验。结果表明,动态阈值云检测方法可用于长时间序列卫星观测数据的云检测,不仅运行速度较快,可实现全自动,且精度较高,陆地上空云检测准确率为82.8%(地面报资料),海洋上空云检测准确率为91.5%(船舶报资料)。此外,还利用CloudSat资料检验不同观测时间差对NOAA系列卫星云检测结果检验的影响。As one of the most successful and extensive polar-orbiting meteorological satellites, NOAA data are widely used in the studies of weather, vegetation index, duststorms and so on. Detection of the cloud presence is the first step of these studies, and its accuracy will directly affect the accuracy of quanti- tative analyses. Almost twenty years from 1988 to 2008, the advanced high resolution radiometer （AVHRR） raw data were collected by China National Satellite Meteorological Center, and their cloud de- tection output by an automated dynamic threshold cloud-masking algorithm have been compared with coin- cident surface weather observations, ship＇s report. The results show that NOAA/AVHRR cloud detection samples are about 82. 8% according with surface weather observation, and approximately 91. 5% with ship＇s report. In addition, CloudSat data are also used to investigate the influence of difference in observa- tion time on the accuracy of NOAA/AVHRR cloud detection.

关 键 词：云检测 NOAA/AVHRR 动态阈值法 

分 类 号：P414.4[天文地球—大气科学及气象学]