Dimensionality reduction on cloud images based on various climate zones

Abstract

In recent decades, cloud image classification has become a research hotspot in the field of weather forecasting. Initially, cloud images that fall on various climate zones are categorized based on their regions. Dimensionality reduction is performed in the cloud images by applying Principal Components Analysis (PCA) to enhance the classification accuracy of cloud images. The proposed system uses the training set to learn the features of cloud images and classifies the test case images into low, medium and high. The experimental results are obtained by implementing the INSAT weather image data set using MATLAB tool. The proposed methodology can be used in various applications like Rainfall Prediction, Oceanography and Cyclone Forecasting.