Institute of Geography
Topic
Automated landslides detection based on airborne laser relief scanning data
PhD. program
Physical Geography, Geoecology and Geoinformatics
Year of admission
2024
Name of the supervisor
Mgr. Ján Novotný, PhD.
Contact:
Receiving school
Faculty of Natural Sciences, Comenius University, Bratislava
Annotation
The development of information technologies such as the use of artificial intelligence and deep learning models brings new opportunities for identifying natural hazards that directly threaten human life and health and cause economic losses. The aim of this thesis will be to use artificial intelligence tools (machine and deep learning) to analyse accurate, high-resolution digital elevation models and test the possibilities of automated slope deformation detection based on texture and morphological image analysis. A detailed statistical analysis will evaluate the suitability of using different combinations of input layers (derived from the elevation model) for identifying individual landslides using pixel-based and object-based classification methods. The work results in the development of a tool for mapping and monitoring these natural hazards.