High resolution (HR) images are of interest in many fields such as medical imaging, surveillance, and satellite imaging. Signal processing based resolution enhancement methods have been finding increasing use due to the limitation...
ver más
¿Tienes un proyecto y buscas un partner? Gracias a nuestro motor inteligente podemos recomendarte los mejores socios y ponerte en contacto con ellos. Te lo explicamos en este video
Fecha límite de participación
Sin fecha límite de participación.
Descripción del proyecto
High resolution (HR) images are of interest in many fields such as medical imaging, surveillance, and satellite imaging. Signal processing based resolution enhancement methods have been finding increasing use due to the limitations in the physical principles as well as limitations in optics and electronics technology. However, none of the methods in curent use performs satisfactorily when applied
to accurate registration. Robust registration models and solutions for multiple object motion and occlusions should be incorporated into these methods. We propose using a super-resolution (SR) reconstruction algorithm based on least squares (LS) filtering combined with a spatiotemporal classification algorithm. Using these classified images
with corresponding LS filter modes and classes, we intend to generate higher resolution images. Registration and multiple object movements
will be incorporated into different LS filter classes without requiring accurate estimation of these changes. The proposed technique will be demonstrated on surveillance images acquired from thermal imaging sensors.