The team is working with an underwater hyperspectral imaging (UHI) camera to map seabed habitat and assess its ecological status. The hyperspectral sensor is a push-broom that records lines of 1,900 pixels perpendicular to the track direction. Intensities of reflected light are continuously measured for spectral bands between 378 and 800 nm with a spectral resolution of 4 nm. Recorded lines need thus to be mosaicked to provide a 2D images. The seafloor is illuminated using LED artificial light.
Yes! Hyperspectral data were acquired during the Chereef cruise in August 2022 in the Lampaul canyon, located off the Brittany Coast. This site is characterised by the presence of great coral habitat diversity and a wide variety of geomorphological features. The UHI was mounted on the Victor IFREMER Hybrid Remotely Operated Vehicle (HROV). A total of 29 profiles from 45 to 90 m long were acquired at 1 to 2 m altitude in a flat area at about 750 m depth. A longer profile of 3,000 m was acquired on the eastern interfluve of the Lampaul Canyon in depths varying from 700 to 1,000 m. In addition, several sediment and species samples were taken to characterise their spectral signatures out of water by carrying out measurements in the laboratory on board. The analysis on the acquired hyperspectral data is currently ongoing to extract key information. The targeted objective is to enhance the knowledge of deep-sea benthic ecosystems as well as assess their ecological status.
Based on passive sensors installed either on aerial platforms or satellites, hyperspectral sensors have been successfully used to detect a wide range of objects and situations in deep and dark environments around the world. While the human eye and conventional optical cameras interpret colours in the Red-Green-Blue (RGB) visible spectral bands, the light is seen through the entire electromagnetic spectrum with hyperspectral vision. This allows capturing views of further details, thus enabling the scientific analysis of object structure through a highly accurate image. Hence, the team decided to extend the application of this technology to deep-sea areas. Up to now, they have developed a processing chain to improve the geometric quality of the UHI images. By combing techniques from simultaneous localisation and mapping, structure-from-motion and 3D reconstruction, they advantageously use them to create 3D models with hyperspectral texture, thus overcoming the flat surface assumption and the classical limitation of dead-reckoning navigation.
Contact: Touria Bajjouk (Touria.Bajjouk[at]ifremer.fr )
Main contributors: Touria Bajjouk, Maxime Ferrera, Tristan Petit, Aurelien Arnaubec, based at IFREMER (France)
© 2024 iAtlantic. All rights reserved.
This project received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 818123 (iAtlantic). This output reflects only the author’s view and the European Union cannot be held responsible for any use that may be made of the information contained therein.