Metodologías para la selección y extracción de características en la mejora de la clasificación de datos hiperespectrales.




MAIN RESEARCHER
Dr. José Martínez Sotoca

Funding by:Consellería d'Educació i Ciencia. Generalitat Valenciana.

Code:
GV/2007/105

Period:
2007-2008



ABSTRACT

The project are framed in Pattern Recognition and Computer Vision Area. So, the spectral information can be very important to improve the description and knowledge in certain tasks. This systems contain spatial and time information captured by spectral sensors from electromagnetic spectrum.

Between the applications where these methodologies can be applied, we must point out: remote sensing, astronomy images, quality assessment in fruits, medical spectral images and so on.

The principal goal is processing the hyperspectral information to find methodologies to obtain better spatial representation in the automatic classification or analyze the knowledge of the different regions in spectral images.