Image Feature Extraction Using Fuzzy Morphology

ISBN-13: 9783639116250 / Angielski / Miękka / 2009 / 124 str.

Elvis Ljumi

Opis

This work proposes a process for extracting featuresfrom remotely sensed imagery that utilizes fuzzymathematical morphology operators. A key concept isthe introduction of an Adaptive Fuzzy Morphology. TheAdaptive Fuzzy Morphology uses an image generatedparameter that is uniquely specific to an image basedon certain image characteristics. That parameter issubsequently used with a set of predefined T-normsand S-norms to generate morphological operators whichprocess the image. Furthermore, the work illustratesmorphological operators based on grayscalemorphology, classical fuzzy morphology, additivemorphology, bounded morphology and adaptivemorphology, and explains the process for defining howthe process of generating the adaptive parameter andthe Adaptive morphological operator works. Thedifferent morphologies are compared on sampleLandsat-7, SPOT-5, QuickBird and EROS images. Theperformance of the Adaptive Fuzzy Morphology iscompared to the other morphologies and evaluated.

O autorze

Elvis Ljumi , Ph.D., received his doctorate in Systems Science
from the State University of New York at Binghamton''s Watson
School of engineering. His areas of specialty are image
processing, remote sensing and fuzzy mathematical morphology. His
background includes computer programming, network administration
and Geographical Information Systems.