Presentation + Paper
19 October 2023 Deep spectral imaging (DSI): an optimized AI-based spectral imaging acquisition, transmission, and calibration
Marcello Melis M.D., Enrico Saggese, Marco Muzi, Matteo Fois, Damiano Scalabrin
Author Affiliations +
Abstract
The evolution of UAV and the great development of micro-satellites pose a number of challenging problems all related to minimize the needed resource on board and to maximize the amount of information obtained from the remote and close sensing activities. What usually happens is to have quite specialized sensors onboard depending on the most required characteristic of the images, high spatial resolution, high spectral resolution, high radiometric accuracy, etc. The authors developed an advanced approach able of minimizing the resources on board while maximizing the amount of information obtained out of them applying AI base post-processing. The on-board equipment and the corresponding ground post-processing provide Spectral and Colorimetric Imaging with notable spectral resolution, accuracy and precision starting from only two acquired images. The result is a Spectral and Colorimetric Imaging System with 13 bands per pixel with precision in the order of 98%, and CIELAB colorimetric coordinates per pixel with a precision of dE2000 ⪅2, starting from only two RAW images.
Conference Presentation
(2023) Published by SPIE. Downloading of the abstract is permitted for personal use only.
Marcello Melis M.D., Enrico Saggese, Marco Muzi, Matteo Fois, and Damiano Scalabrin "Deep spectral imaging (DSI): an optimized AI-based spectral imaging acquisition, transmission, and calibration", Proc. SPIE 12729, Sensors, Systems, and Next-Generation Satellites XXVII, 127291D (19 October 2023); https://doi.org/10.1117/12.2679981
Advertisement
Advertisement
RIGHTS & PERMISSIONS
Get copyright permission  Get copyright permission on Copyright Marketplace
KEYWORDS
Tunable filters

Optical filters

Artificial intelligence

Image filtering

Sensors

Calibration

Imaging spectroscopy

Back to Top