极轨气象卫星图像的云自动检测方法研究

Automatic Cloud Detection of Polar Orbit Meteorological Satellite Image

图像上云的存在给遥感图像的处理与分析带来了不便。本文探讨了利用极轨气象卫星图像可见光波段的反射率和热红外波段的亮温信息进行云自动检测的方法。该方法在可见光波段是否有效判断的基础上,进行图像水陆区域划分,然后对不同图像、不同波段的水域和陆域分别设定合理的阈值,实现了云自动检测目的。试验结果表明,该方法能够取得较高的云检测精度。

Cloud in image is the obstacle for processing and analyzing of remote sensing images. The paper discussed a method which automatically detect cloud in polar orbit meteorological satellite images by using visual band reflectance and thermal infrared bright temperature. Based on judge whether the visual band was valid or not, the method separated land region from water region, and then set proper threshold values separately for land and water regions in different images and different bands, by that way an automatic cloud detection could be achieved. Experimental results showed that the method can get a higher precision for cloud detection.