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 utilizes 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 proposed Adaptive Fuzzy Morphology is compared to the other morphologies and evaluated.