SATELLITE IMAGE CLASSIFICATION OF WATER, ICE, AND LAND USING DEEP LEARNING

Abstract

Satellite image classification plays a crucial role in environmental monitoring, climate analysis, and resource management. The ability to accurately distinguish between water, ice, and land surfaces is essential for applications such as glacier tracking, flood monitoring, and land-use planning. Traditional image processing techniques often struggle with spectral similarities and noise in satellite imagery. To address these challenges, deep learning methods, particularly Convolutional Neural Networks (CNNs), have emerged as powerful tools for automated image classification. This paper presents a deep learning-based approach for classifying satellite images into three primary categories: water, ice, and land. The proposed model leverages high-resolution satellite datasets and applies preprocessing techniques such as normalization, resizing, and augmentation to enhance model performance. A CNN architecture is designed to extract spatial features and learn complex patterns that differentiate the three classes effectively.