NOAA AVHRR云标识技术的应用分析

中国农情遥感速报系统对每日的NOAA AVHRR进行定量处理，其中用CLAVR方法进行云标识。本文选择三景不同地区的NOAA AVHRR影像，分析评价CLAVR方法云标识的效果和各个步骤的标识能力，以及空间上的差异，并根据运行经验，对CLAVR方法的部分参数进行了调整，使其更加适应中国大陆的情况。总体上，CLAVR方法标识干净像元和云污染像元的准确性较高，而在标识混合像元时稍差。其中RGCT、RUT、TUT和C3AT的检出率占总检出率89．2％以上，并在不同的地区，不同步骤的贡献度不一致，也说明了云相变化随区域不同而变化。

Application of Cloud Detection Algorithm for the AVHRR Data

Application and analysis of the algorithm calles CLAVR were made for thr remote sensing of cloud cover using multispectral radiance measurements from the Advanced Very High Resolution Radiometer(AVHRR)on board National Oceanic and Atmospheric Administration(NOAA) polar-orbiting satellites.The algorithm classifies 2*2 pixel arrays from the HRPT 1.1-km resolution data into CLEAR,MIXED,and CLOUDY categories.It uses a sequence of multispectral contrast,spectral,and spatial signature threshold tedts to perform the classification.The algorithms on China land area were applied to realtimg NOAA-14 data and about 10 years dataset of China land area have been processed.It did well in classifying perfectly clear and cloudy pixel arrays except existing some errors for mixed pixel arrays.Different images in China land area show some difference,and the comolicacy of landform and underlying surface influences cloud's appearance or structure,thereby resulting contrast for different steps of the algorithm.The thresholda of main steps (RRCT,TUT and RUT)have been modified in the algorithm that required in the application of decision of decision tree tests.It retrieves some CLOUDY and CLEAR pixels from the original classified MIXED pixels,sepaeately the 4.13% and 1.09%.The CLAVR has been integrated into the Crop Growing Monitioring system of China.