Morphological segmentation of hyperspectral images

TL;DR

This paper introduces a watershed-based morphological segmentation method for hyperspectral images, combining spectral classification, vectorial gradients, and data reduction techniques to improve segmentation accuracy across various spectral spaces.

Contribution

It presents a novel general methodology for hyperspectral image segmentation that integrates spectral classification, multiple gradient adaptations, and data reduction methods.

Findings

Effective segmentation across different spectral spaces

Improved spatial and spectral information integration

Versatile approach applicable to various hyperspectral datasets

Abstract

The present paper develops a general methodology for the morphological segmentation of hyperspectral images, i.e., with an important number of channels. This approach, based on watershed, is composed of a spectral classification to obtain the markers and a vectorial gradient which gives the spatial information. Several alternative gradients are adapted to the different hyperspectral functions. Data reduction is performed either by Factor Analysis or by model fitting. Image segmentation is done on different spaces: factor space, parameters space, etc. On all these spaces the spatial/spectral segmentation approach is applied, leading to relevant results on the image.