Fortschritte und Herausforderungen bei der Erkennung und Klassifizierung von vulkanischer Asche mithilfe von satellitengestützten UV/VIS-Messungen und Deep Learning
Abstract
This study introduces a new method to retrieve volcanic ash and its microphysical properties from ultraviolet and visible satellite data. Neural networks trained on simulated radiative transfer data achieved over 90% accuracy in distinguishing volcanic ash from other aerosols and clean atmospheres. The method was applied to satellite measurements, demonstrating its ability to identify volcanic ash and estimate its optical and microphysical properties.
Fields of Study & Subjects
Atmospheric aerosols and clouds
Atmospheric Ozone and Climate
Remote Sensing in Agriculture
Publication details
- Journal / source
- mediaTUM (Technical University of Munich)
- Publication date
- 1 January 2026
- Language
- EN
- Document type
- Thesis