From sky to seafloor observation: Achieving eXcellence in Oceanic surveiLlance and cOnservation Through deep Learning

Descripción del proyecto

Mejorar la conservación marina y la vigilancia marítima mediante el aprendizaje profundo

El equipo del proyecto AXOLOTL, financiado con fondos europeos, tiene como objetivo mejorar la capacidad de investigación y desarrollo del Instituto Marítimo y Marino de Chipre (CMMI, por sus siglas en inglés) en materia de evaluación de la biodiversidad marina y vigilancia marítima utilizando tecnologías punteras de aprendizaje profundo. Los modelos de aprendizaje profundo se entrenarán utilizando conjuntos de datos estandarizados y robustos de datos muestreados «in situ» y datos de teledetección procedentes de satélites y acústica submarina. Las soluciones propuestas se validarán en escenarios reales en torno a la evaluación de la biodiversidad y la vigilancia marítima en contextos chipriota, belga y europeo. El equipo del proyecto también tratará de tender un puente entre el CMMI, un centro de excelencia de trabajo en equipo, y los investigadores e innovadores de Bélgica y Francia de prestigio internacional. Para ello, facilitará la transferencia de conocimientos, la creación de redes para la excelencia científica y las competencias administrativas y empresariales.

Objetivo

Technological areas such as Artificial Intelligence (AI) and ecosystems such as shipping, maritime and space have been strategically prioritised by Cyprus to improve its Research and Innovation (R&I) performance.

At their intersection lies the need to enhance marine conservation efforts and maritime surveillance by leveraging deep learning (DL) methodologies built on standardized and robust data sets. Indeed, in-situ sampling stands as a cornerstone in marine conservation, offering a direct approach to monitoring marine biodiversity; while remote sensing stands as a pivotal addition to maritime surveillance, expanding the scope beyond traditional Automatic Identification System capabilities. DL, as a cutting-edge AI tool, holds immense potential to enhance the analysis of in-situ samples and remotely sensed data.

AXOLOTL is proposed as a transformational international endeavour capable of enhancing the R&I capacity of CMMI, Cyprus and Europe in the interdisciplinary fields of DL-enhanced in-situ biodiversity assessment and DL-enhanced remote sensing for maritime surveillance. The project will contribute to closing the gap between a H2020 Teaming Centre of excellence and 2 strong innovators from France and Belgium through capacity-building, knowledge transfer, networking, and outreach activities at regional and international levels. Activities will go beyond the strictly scientific scope and support the mutual development, consolidation, and reinforcement of administrative, dissemination and entrepreneurial competencies, access to networks of excellence and the sustainable linkage between partners. The project’s R&I component will develop new strategies for improving data quality, standardization, and synchronization issues, devise novel interdisciplinary methodologies, develop robust DL models from state-of-the-art computer vision methods, and validate its proposed solutions in 2 relevant real-world contexts (biodiversity assessment and maritime surveillance).