AbstractMorphological and Attribute Profiles for Classification of Hyperspectral Images About a decade has passed since the concept of morphological profile (MP) was defined for the analysis of panchromatic remote sensing images. From that time, the MP has largely proved to be a powerful tool able to model spatial information (e.g., contextual relations) of the image by extracting structural features (e.g., size, geometry, etc.) from the objects present in the scene. The MP processes an input image with a sequence of progressively coarser filters. This leads to a stack of filtered images showing an increasing simplification of the scene. The evaluation of how the objects in the image interact with the filters gives information on the objects structural features. The great amount of contributions present in the literature that address the application of MP to many tasks (e.g., classification, object detection, segmentation, change detection, etc.) and to different types of images (e.g., panchromatic, multispectral, hyperspectral) proves how MP is still an effective and modern tool. Moreover, many variants, extensions and refinements of its definition have also appeared stating that the MP is still under continuous development. In the talk the MP is presented from its early definition to the recent advances based on morphological attribute filters. The morphological attribute filters are used in classification hyperspectral images and the obtained results are compared to results obtained with no morphological processing. |