International Journal of Engineering and Technology

Clouds are important for climatic changes in the atmosphere. Cloud images are taken by visible and infrared satellites. Both visible and infrared satellites, the clouds that are not very white are called Thin Clouds. So, the identification of a thin cloud is hard to find out. In this paper, some algorithms for spotting thin clouds are proposed. They are HSL (Hue Saturation Light), ROI (Region of Interest), Watershed, Demirel and Color Segmentation. HSL (Hue Saturation Light) is used for finding the ratio of color in the cloud image, ROI (Region of Interest) is used for removing the cloud and the sky elements of the cloud image, Watershed and Demirel algorithm are used to segment the Thin Clouds. The extract values from the cloud image are measured using Gaussian Filter. The weather forecasting is carried out by comparing the images in the database after performing color segmentation by k-means clustering.