RESEARCH ON AUTOMATIC FOG IDENTIFICATION TECHNOLOGY BY METEOROLOGICAL SATELLITE REMOTE SENSING

There is an urgent need for the development of a method that can undertake rapid, effective, and accurate monitoring and identification of fog by satellite remote sensing, since heavy fog can cause enormous disasters to China’s national economy and people's lives and property in the urban and coastal areas. In this paper, the correlative relationship between the reflectivity of land surface and clouds in different time phases is found, based on the analysis of the radiative and satellite-based spectral characteristics of fog. Through calculation and analyses of the relative variability of the reflectivity in the images, the threshold to identify quasi-fog areas is generated automatically. Furthermore, using the technique of quick image run-length encoding, and in combination with such practical methods as analyzing texture and shape fractures, smoothness, and template characteristics, the automatic identification of fog and fog-cloud separation using meteorological satellite remote sensing images are studied, with good results in application.