基于多种底图的云检测  

Cloud detection based on bottom image

作　　者：谭永强[1] 费建芳[1] 石汉青[1] 

机构地区：[1]解放军理工大学气象学院,南京211101

出　　处：《气象科学》2009年第1期52-57,共6页Journal of the Meteorological Sciences

基　　金：中国科学院大气物理研究所LASG实验室开放课题(2709);中国气象局上海台风研究所开放课题(2006STB02)

摘　　要：利用静止卫星红外和短波红外通道资料的特点,生成了红外亮温底图与红外减短波红外差异底图,并充分利用这几种底图的特性,结合时空相似性检测来进行云检测的校正,实现了用动态阈值技术进行云检测。分析表明,红外亮温底图对于中高云系的检测效果较好,差异底图对于中低云系的检测效果较好,通过时空不同性测试可以将误检为云的晴空重新识别出来。算法取得了较好的检测效果。The characteristics of infrared and near infrared channel data in geostationary satellite have been used to generate the bottom image of infrared brightness temperature and infrared subtract near infrared brightness temperature. Then these bottom image and the method of resemblance in time and space are able to correct the results and achieve the goal of using different thresholds in cloud detection. The detection results showed that the infrared bottom image is very effective on detection of high and medium height cloud, and the bottom image of infrared subtract near infrared brightness temperature is very effective on detection of low height cloud. Using method of resemblance in time and space can correct the cloud detection result. This algorithm got a good result.

关 键 词：云检测 短波红外通道 底图 相似性 

分 类 号：P455[天文地球—大气科学及气象学]