types of typical meteorological cloud, land and sea water were selected in the study. Next, taking the MODIS bands 30 and 31 as an example, the unsupervised classification of two-dimensional spectral feature space was carried out by fuzzy clustering remote sensing classification method with neighborhood distance constraint. Finally, from the practical point of view, five typical types of meteorological clouds were selected from the common meteorological cloud sample database that were used to fuzzy clustering classified in two-dimensional spectral feature space, and further got the cluster classification area of volcanic ash cloud and other types of meteorological cloud in two-dimensional spectral feature space. Figure 8 illustrates the results of four sample area‘s fuzzy clusters in MODIS bands 30 and 31 by the proposed fuzzy clustering remote sensing classification method with neighborhood distance constraint in this paper.
Based on the classification experiment by fuzzy clustering remote sensing classification method with neighborhood distance constraint, in order to more clearly illustrate the volcanic ash cloud after the classification, the five types of classified meteorological clouds, i.e., cumulonimbus cloud, cirrus cloud, medium cloud, low cloud, and cumulus cloud, are merged into one type of meteorological cloud. The classification result by the proposed method is show in Fig. 9. As shown in Fig. 9, the distribution shape of classified volcanic ash cloud is basically consistent with visual interpretation.
F