Cloud Classification from NOAA Satellite Image Using Learning Vector Quantization Method

Abstract

Cloud images from NOAA satellites 18 and 19 are essential for weather forecasting and climate analysis. Imagery from satellites in the cloud’s shape can be distinguished based on the cloud (low, middle, and high). This paper uses the multilevel thresholding segmentation method compared with the FCM method (fuzzy c-mean clustering). The segmented data with the two methods are classified using the LVQ method. This study’s results obtained the accuracy of the cloud data recognition segmented using multilevel thresholding of 72.22% and cloud data segmented using FCM of 83.33%.