Classification and change detection in multi - epoch airborne laser scanning point clouds

Abstract

Thesis introduces data organization and pre-processing procedures, including pyramid generation and data filtering; includes a literature review of algorithms suitable for the classification of lidar data; explains why they are included and how the choice of entity is made and combined with the method; includes an evaluation and analysis of the classification, including accuracy, confusion matrices, impact of features, and the impact of multiple entities; reviews the state-of-the-art of change detection with lidar, and covers the change detection methodology used in my research; involves examples of all types of change in buildings and a list of errors and the reasons for their occurrence.