POSTER SESSION 1
15:30–17:30, Monday, September 14
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
ABSTRACT 1247 | POSTER 141
OCEAN COLOR SUPER-RESOLUTION IMAGES USING BLIND FUSION OF SENTINEL-2 AND SENTINEL-3 DATA
Coastal environments represent a dynamic interface between terrestrial and marine ecosystems, hosting nearly 60% of the global population and facing major ecological, economic, and societal challenges. These areas are particularly vulnerable to health risks such as harmful algal blooms, as well as to the impacts of climate change. Ocean color is a key indicator for assessing water quality and monitoring marine ecosystem dynamics. In this context, satellite remote sensing provides an effective solution, offering large-scale coverage and frequent revisit capabilities. However, coastal observation requires both high spatial and high spectral resolution, provided respectively by the MSI sensor onboard Sentinel-2 and the OLCI sensor onboard Sentinel-3. Due to inherent physical limitations, these two properties cannot be achieved simultaneously by a single sensor. Conventional fusion approaches typically rely on combining only two images, which prevents full exploitation of multi-resolution data, particularly from Sentinel-2. To address this limitation, we previously introduced two generalized methods, G-STEREO-1 and G-STEREO-2, based on a tensor framework relying on Canonical Polyadic Decomposition (CPD), enabling the joint fusion of multiple images. While these methods demonstrated strong performance compared to state-of-the-art approaches, they rely on the assumption that spatial degradation operators are known, which is rarely satisfied in practice. In this work, we propose a blind extension, BG-STEREO, which removes this assumption by implicitly estimating the degradation operators within the tensor model. Experimental results on both simulated and real data demonstrate that the proposed method maintains high reconstruction accuracy while improving robustness under realistic acquisition conditions and diverse environmental scenarios.
Fatima Zahrae M’Rani*, Université du Littoral Côte d’Opale, [email protected]
Matthieu Puigt, Université du Littoral Côte d’Opale, [email protected], https://orcid.org/0000-0002-3264-4981
Cédric Jamet, Université du Littoral Côte d’Opale, [email protected], https://orcid.org/0000-0001-6988-6506
Claire Guilloteau, Université du Littoral Côte d’Opale, [email protected]
Gilles Delmaire, Université du Littoral Côte d’Opale, [email protected]
Gilles Roussel, Université du Littoral Côte d’Opale, [email protected]
Vincent Vantrepotte, Université du Littoral Côte d’Opale, [email protected]
Poster Session | 1 | 2 | 3 | 4 | Schedule at a Glance
