Image sequences play an important role in today's world. They provide us a lot of information. Videos are for example used for traffic observations, surveillance systems, autonomous navigation and so on. Due to bad acquisition, transmission or recording, the sequences are however usually corrupted by noise, which hampers the functioning of many image processing techniques. A preprocessing module to filter the images often becomes necessary. After an introduction to fuzzy set theory and image processing, in the first main part of the thesis, several fuzzy logic based video filters are proposed: one filter for grayscale video sequences corrupted by additive Gaussian noise and two color extensions of it and two grayscale filters and one color filter for sequences affected by the random valued impulse noise type. In the second main part of the thesis, interval-valued fuzzy mathematical morphology is studied. Mathematical morphology is a theory intended for the analysis of spatial structures that has found application in e.g. edge detection, object recognition, pattern recognition, image segmentation, image magnification… In the thesis, an overview is given of the evolution from binary mathematical morphology over the different grayscale morphology theories to interval-valued fuzzy mathematical morphology and the interval-valued image model. Additionally, the basic properties of the interval-valued fuzzy morphological operators are investigated. Next, also the decomposition of the interval-valued fuzzy morphological operators is investigated. We investigate the relationship between the cut of the result of such operator applied on an interval-valued image and structuring element and the result of the corresponding binary operator applied on the cut of the image and structuring element. These results are first of all interesting because they provide a link between interval-valued fuzzy mathematical morphology and binary mathematical morphology, but such conversion into binary operators also reduces the computation. ...
Fuzzy techniques for noise removal in image sequences and interval-valued fuzzy mathematical morphology
2010-01-01
Hochschulschrift
Elektronische Ressource
Englisch
DDC: | 629 |
British Library Online Contents | 2012
|Random impulse noise removal from image sequences based on fuzzy logic
British Library Online Contents | 2011
|On the Decomposition of Interval-Valued Fuzzy Morphological Operators
British Library Online Contents | 2010
|