Analysis of Remotely Sensed Imagery Using Digital Morphology

摘要

An approach to the automated extraction of cultural terrain features from remotely sensed imagery is the development of structuring elements that are based on the concepts of digital morphology. The image structures of cultural terrain features show a great deal of regularity and can be represented by models. Operators that measure and analyze image structures provide effective tools for the automated feature extraction process. Morphological operations that belong to the class of hit-or-miss transforms are discussed in some detail. The operations of dilation and erosion are explained and demonstrated on radar imagery. An operation that combines gray-tone differencing and thresholding is used for extracting a road segment from an image. The future applications of digital morphology for analyzing and measuring image primitives, attributes, descriptors and descriptor sets of terrain features in support of automated feature extraction are briefly discussed.