PROJECT TITLE :

Vector Attribute Profiles for Hyperspectral Image Classification

ABSTRACT:

Morphological attribute profiles are among the most prominent spectral–spatial pixel description ways. They are economical, effective, and highly customizable multiscale tools based on hierarchical representations of a scalar input image. Their application to multivariate pictures generally and hyperspectral pictures in explicit has been so far conducted using the marginal strategy, i.e., by processing every image band (eventually obtained through a dimension reduction technique) independently. During this paper, we investigate the choice vector strategy, which consists in processing the out there image bands simultaneously. The vector strategy relies on a vector-ordering relation that results in the computation of one max and min tree per hyperspectral data set, from which attribute profiles can then be computed as usual. We tend to explore known vector-ordering relations for constructing such max trees and, subsequently, vector attribute profiles and introduce a combination of marginal and vector ways. We offer an experimental comparison of these approaches in the context of hyperspectral classification with common information sets, where the proposed approach outperforms the widely used marginal strategy.