Artificial Intelligence Methods for Classifying and Detecting Changes in Remote Sensing Data

This research focuses on the use of artificial intelligence methods for the classification and detection of changes in remote sensing data. In recent years, there has been a significant increase in the volume of satellite and aerial data with high spatial, spectral, and temporal resolution, which places high demands on automated methods for their processing and analysis. The work focuses primarily on modern machine learning and deep learning approaches, such as convolutional neural networks, recurrent neural networks, and unsupervised learning methods, and their application to multispectral, hyperspectral, and time-series ESR data.
The main objective is to analyze and compare selected artificial intelligence methods in terms of accuracy, robustness, and computational complexity for tasks involving land cover classification and the automatic detection of land-use changes, for example in connection with urbanization, deforestation, agricultural land use, or the impacts of natural disasters. The topic also addresses the issues of training data preparation, model transferability across different regions and sensors, and the interpretability of results. The results may contribute to more effective landscape monitoring and support decision-making in the areas of land-use planning, environmental protection, and crisis management.