| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 270.88 KB | Adobe PDF |
Advisor(s)
Abstract(s)
Hyperspectral imaging instruments allow remote Earth exploration by measuring hundreds of spectral bands at very narrow channels of a given spatial area. The resulting hyperspectral data cube typically comprises several gigabytes. Such extremely large volumes of data introduces problems in its transmission to Earth due to limited communication bandwidth. As a result, the applicability of data compression techniques to hyperspectral images have received increasing attention. This paper, presents a study of the power and time consumption of a parallel implementation for a spectral compressive acquisition method on a Jetson TX2 platform. The conducted experiments have been performed to demonstrate the applicability of these methods for onboard processing. The results show that by using this low energy consumption GPU and integer data type is it possible to obtain real-time performance with a very limited power requirement while maintaining the methods accuracy.
Description
Keywords
Hyperspectral imagery Compressive sensing Embedded systems
Citation
NASCIMENTO, José M. P.; VÉSTIAS, Mário – Low power compressive sensing for hyperspectral imagery. In 11th Conference on Telecommunications – ConfTele 2019. Sede da Ordem dos Engenheiro – Lisboa (Portugal), June 26-28th, 2019