Identification of Fog with NOAA AVHRR Images

Fog often occurs from June to September on the Pacific side of northern Japan. As a consequence of poor visibility, there are many disasters at sea. By identifying fog using satellite data, accidents at sea could be reduced. In this paper, we attempted to identify fog using image-processing data from the NOAA Advanced Very High Resolution Radiometer (AVHRR). AVHRR sensors measure radiation at visible, near-infrared, and infrared wavelengths. Here we restrict our attention to daytime data from channel 1 (CHI), channel 3 (CH3), and channel 4 (CH4) for a detection range of λ = 0.58 ~ 0.68 μm, λ = 3.55 ~ 3.93 μm, and λ= 10.5 ~ 11.5 μm, respectively. We define the fog in this paper as observed by a lighthouse keeper. We use CH3 and CH4 because these channels contain the thermal information of such fog. We have analyzed by image processing, the data obtained from CHI, CH3, and CH4, by forming two- and three-dimensional histograms and enhancing them in color. The fog analysis technique enables a nonmeterologist to have a preliminary guess regarding the regions of fog in this histogram.