Four-dimensional histogram method for sea ice detection using NOAA AVHRR images

In Far East Russia, the Sea of Okhotsk is covered by sea ice in winter every year. The Amur River plays the important role in generating sea ice. Sea ice goes south by the monsoon. We have developed four-dimensional histogram method for sea ice detection using NOAA AVHRR images. The data in February to March is used for five years from 1995 to 1999. The data of 179 scenes is used among 699 scenes to create four-dimensional histogram. We applied the four-dimensional histogram to detect sea ice of the Sakhalin area. As a result of applying the presented method, thick ice was undetectable although thin ice was detectable. Since thin ice is generated by the continental area along the shore, it is effective in observations of this area. Moreover, in the detection result, clouds and sea ice were distinguishable. Furthermore, when comparison with the detection result by the oceanography researcher was performed, the domain detected by this method was in agreement with the detection result by the researcher. This is the next subject although it is necessary to increase the number of scenes in order to detect thick ice.