A technique for simultaneous visualization and segmentation of hyperspectral data
Download as DOCX, PDF0 likes167 views
This document proposes a method for simultaneously visualizing and segmenting hyperspectral data. The method treats the hyperspectral data set as a single image created by weighted combination of spectral points for each pixel, with the weights optimized to improve human visibility. A piecewise-constant smoothness constraint on the weights prevents blurring in the fused image. Optimizing the weight matrix also provides information to spatially segment the original hyperspectral data set. Experimental results demonstrate the effectiveness of the proposed method.
1 of 1
Download to read offline
More Related Content
What's hot (9)
Ad
Viewers also liked (6)
Ad
Similar to A technique for simultaneous visualization and segmentation of hyperspectral data (20)
More from I3E Technologies (20)
A technique for simultaneous visualization and segmentation of hyperspectral data
- 1. A TECHNIQUE FOR SIMULTANEOUS VISUALIZATION AND SEGMENTATION OF HYPERSPECTRAL DATA ABSTRACT In this paper, we propose an optimization-basedmethod for simultaneous fusion and unsupervised segmentationof hyper spectral remote sensing images by exploiting redundancyin the data. The hyperspectral data set is visualized as a singleimage obtained by weighted addition of all spectral points at eachpixel location in the data set. The weights are optimized to improvethose statistical characteristics of the fused image, which invoke anenhanced response from a human observer. A piecewise-constantsmoothness constraint is imposed on the weights instead of thefused image by minimization of its 3-D total-variation norm, thuspreventing the fused image from blurring. The optimal recoveryof the weight matrix additionally provides useful information insegmenting the hyperspectral data set spatially.We provide ampleexperimental results to substantiate the usefulness of the proposedmethod.