A Fuzzy Logic Approach for Cloud Classification based on Near-Infrared Image Features

The cloud type classification and analysis from satellite image have been topic of research interests in many atmospheric and environmental studies but still is a challenging task. In this paper, we propose a new method for cloud type classification using fuzzy logic. We include near-infrared images as an input source in addition to visible and infrared in our previous study. Using that information source, we can successfully improve the noise removal process and the firsthand classification with respect to the height of cloud existence by considering reflection and release characteristics. Two fuzzy membership functions are designed to make final classification decision. In experiment, the proposed method is verified to be efficient and more accurate than the previous attempt that mainly used infrared image features.