Cloud Extraction from INSAT-3D Satellite Image by K-Means and Fuzzy C-Means Clustering Algorithms

This paper presents algorithms for extraction of clouds from INSAT-3D satellite image over the Indian region. The K-Means and Fuzzy C-Means clustering algorithms are applied on INSAT-3D satellite images on some specific dates and time in the year 2017, when the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite crosses the Indian region. Prior to this, the number of cluster segments k is selected from the MODIS Aqua sensor's cloud product. The result of segmentation algorithms is validated by comparing with the cloud optical thickness of the MODIS data. The comparison shows that INSAT-3D cloud segmentation matches well with the cloud optical thickness of the MODIS data.