MODIS数据的云检测处理

Cloud detection in MODIS data

结合MODIS数据的通道特性,分析了云在不同波谱范围的光谱特性,提出了一种基于多光谱综合阈值法的简便新方法。该云检测算法从可见光反射率、热红外亮温出发,逐步建立云检测模板,实现分离云和晴空,通过对不同下垫面类型数据进行云检测检验,结果表明,该算法对中低纬度地区云检测处理有显著效果,尤其能够检测出可见光波段难以识别的薄卷云。该法可使遥感数据特别是MODIS数据在云检测领域的应用提高一个层次,进而提出的云区数据修补方法,使数据反演应用的精度提高更为显著。

Through analyzing the meteoric characteristic of cloud in different spectrum and integrating the spectrum characteristic of MODIS （Moderate resolution imaging spectroradiometer）, a new and simple cloud detection method was put forward which based on multi-spectrum synthesis. This algorithm referred to visible reflection index, infrared brightness temperature and window brightness temperature difference and built a cloud detection mask step by step, fiinally, an entire cloud processing result was found out to seperate cloud from clear-sky. The algorithm was applied to different time and different scenes to validate and analyze. Results show that the cloud detection is ideal for mid or low latitude region. Especially, this algorithm has a good effect for thin-cirrus which is invisible in visible band. On the base of cloud detection, the method of cloud replacement was put forward, which can promote the use of MODIS data and improve the accuracy of retrieving.