Analysis and Generalization of Machine Vision Algorithms with Subpixel Accuracy
Dr. Óscar Reinoso García

This Doctoral Thesis is focused on the study and development of methods that allow to obtain characteristics or properties of digital images with greater accuracy
to obtain characteristics or properties of digital images with greater precision than the resolution of the
resolution of the digital image (subpixel level). The resolution available in the study of digital images has
digital images has always been conditioned by the Image Acquisition System used.
used. The understanding and modeling of this system has allowed a better knowledge of the possibilities of increasing the resolution.
possibilities of increasing such resolution.
 
Edge detection is a crucial and determinant stage in many processes using Computer Vision techniques.
using Computer Vision techniques. Often, the higher the precision with which edges are detected and located
the better the results obtained in subsequent processing stages used.
processing steps employed. Edge detection at the sub-pixel level thus represents an important solution to
This is an important solution to increase the resolution of the initial data.
 
There are many applications in Computer Vision where morphological transformations are used.
morphological transformations. Until now, these have been conditioned, in the discrete field of digital images, by the size of the pixel.
digital images, by the size of the pixel used. The possibility of using
morphological transformations at the supbixel level will allow the solution of hitherto unresolved problems and the
and the improvement of the results obtained by many of the applications that have been developed.