Angličtina
This work proposes a process for extracting features§from remotely sensed imagery that utilizes fuzzy§mathematical morphology operators. A key concept is§the introduction of an Adaptive Fuzzy Morphology. The§Adaptive Fuzzy Morphology uses an image generated§parameter that is uniquely specific to an image based§on certain image characteristics. That parameter is§subsequently used with a set of predefined T-norms§and S-norms to generate morphological operators which§process the image. Furthermore, the work illustrates§morphological operators based on grayscale§morphology, classical fuzzy morphology, additive§morphology, bounded morphology and adaptive§morphology, and explains the process for defining how§the process of generating the adaptive parameter and§the Adaptive morphological operator works. The§different morphologies are compared on sample§Landsat-7, SPOT-5, QuickBird and EROS images. The§performance of the Adaptive Fuzzy Morphology is§compared to the other morphologies and evaluated.
