Cloud Satellite Image Segmentation using Meng Hee Heng K-Means and DBSCAN Clustering

Satellite image segmentation contains a most significant role to play within the field of remote sensing imaging, for detection of the surface of the planet effectively. One of the satellite image that available in Indonesia is Himawari 8 IR enhanced, provided by Indonesian Agency for Meteorology, Climatology and Geophysics, updated every hour. This satellite image provides information about cloud in Indonesia categorized on its temperature and height. In this study, we experiment clustering algorithm as a segmentation technique to detect the cloud form on Himawari 8 image. Meng hee heng k-means and DBSCAN proposed as the algorithm. Both of algorithms give a stable cluster numbers on every data with data range value between 0.45 - 0.47. The comparison result indicates that DBSCAN can obtain more specific result of the cloud form division showed by the number of cluster obtained.