Superpixels for Hyperspectral Image Analysis



Journal Title

Journal ISSN

Volume Title



With rapid development of multi-channel optical imaging sensors, hyperpsectral data has become increasingly popular, necessitating development of algorithms for robust image analysis with such data. This thesis contributes methods that efficiently and robustly exploits superpixels for hyperspectral data. We study and quantify the efficacy of state-of-the-art superpixel generation algorithms for a variety of hyperspectral images. In this work, superpixel level analysis is proposed for two different hyperspectral image analysis problems — remote sensing image classification and person re-identification via forward looking hyperspectral imagery. Specifically, for remote sensing images, we propose a framework based on superpixels that provides spatial context for robust classification, and, for ground-based “natural” hyperspectral images, efficacy and utility of superpixels is demonstrated, in a multi-view setup, through a pilot study on a person re-identification task.



Hyperspectral imaging


Portions of this document appear in: Priya, Tanu, Saurabh Prasad, and Hao Wu. "Superpixels for Spatially Reinforced Bayesian Classification of Hyperspectral Images." IEEE Geosci. Remote Sensing Lett. 12, no. 5 (2015): 1071-1075.