Abstract:

According to the spectral radiation characteristics of clouds, fog and underlying surface and texture features of clouds and fog, combined with the previous studies, a series of discriminant indexes were established by using object-oriented method, firstly. Then, the algorithm was constructed to detect sea fog from FY-2E data in the daytime, which was applied to a dynamic process of the sea fog in the middle and northern Yellow Sea on 8 April 2014. The applying result showed that the proposed algorithm of sea fog detection in this paper could better monitor the dynamic change of the sea fog in the middle and northern Yellow Sea on 8 April 2014. In addition, the accuracy of the algorithm was tested based on eleven times fog product from FY-3B. The test results showed that the probability of detection (POD) was 90.9%, false alarm rate (FAR) was 33.2%, and critical success index (CSI) was 62.6%, which indicated that the proposed method in the paper was effective and feasible.

Key words: FY-2E, sea fog, dynamic threshold, cloud and fog separation

摘要：

根据各类云层、雾和下垫面的光谱辐射特性和云雾的空间纹理，在前人研究基础上，采用面向对象的方法，构建一系列判别指数，建立一套有效的海雾检测算法进行FY-2E数据白天海雾的判识，并将该方法应用到2014年4月8日黄海中北部一次海雾的动态变化过程。2014年4月8日高时间分辨率的海雾检测个例应用表明，本文提出的白天海雾检测算法可较好地实现海雾过程的动态监测。另外，结合FY-3B雾产品数据进行算法的精度检验，11次海雾个例的精度检验结果显示，命中率（POD）为90.9%，误报率（FAR）为33.2%，临界成功指数（CSI）为62.6%，表明该方法有效、可行。

关键词: FY-2E, 海雾, 动态阈值, 云雾分离