Author: A. A. Sirota and A. Yu. Ivankov (Voronezh State University, Voronezh, Russia)
Optoelectronics, Instrumentation and Data Processing, Volume 51, Issue 6, pp 622–631 (November 2015)
DOI: https://doi.org/10.3103/S8756699015060151
Abstract
Algorithms for improving the resolution of a sequence of images distorted by applicative noise manifested in the form of randomly shaped local regions of occlusion are synthesized and analyzed. The basis is the implementation of an optimal linear filtering algorithm in block form, which is synthesized with account for possible emergence of false observations in individual elements of the analyzed images. A method for incorporating the results of independent segmentation of each individual image into the recursive filtering process is proposed, which enhances the quality of the final processing.
Keywords
image processing, segmentation, applicative noise, super resolution, the Kalman filter
Full text