PhD Defense: 'Techniques for the extraction of spatial and spectral information in the supervised classification of hyperspectral imagery for land-cover applications'
The objective of this doctoral thesis is the development of spatial-spectral information extrac-tion techniques for supervised classification tasks, using both classical and deep learning-based models, intended for the classification of remotely sensed multi and hyperspectral, land use, and land cover (LULC) images. The main objective is to apply these techniques in an efficient way so that they are able to obtain satisfactory classification results with low use of compu-tational resources and execution time. Two research lines are developed as part of this thesis: A first research line oriented towards the development of information extraction techniques designed to be used with classical models, as well as their adaptation to multi-core architec-tures and consumer GPUs (Graphics Processing Unit). A second research line oriented to the development of information extraction techniques aimed at data augmentation for models based on deep learning.
Supervisors: Dora Blanco Heras & Francisco Santiago Argüello Pedreira
Virtual event
/events/phd-defense-techniques-for-the-extraction-of-spatial-and-spectral-information-in-the-supervised-classification-of-hyperspectral-imagery-for-land-cover-applications
events_en