Autonomous systems are enabling new approaches to observing marine ecosystems by combining scalable platforms with diverse sensing modalities. This work presents a multi-modal operational framework for marine species detection that leverages complementary approaches, including vision-based imaging, active acoustics, and environmental DNA (eDNA), deployed across autonomous surface and underwater vehicles.
We demonstrate how different sensing modalities are suited to distinct species and monitoring objectives, and how their coordinated use improves detection robustness, spatial coverage, and ecological insight. Each modality is applied in a targeted manner depending on species characteristics and observability constraints. Vision-based methods support fine-scale identification of species such as handfish. Active acoustics enable monitoring of species such as snapper over extended spatial scale. eDNA sampling provides a non-invasive mechanism for detecting cryptic or transient species that may not be observable through visual or acoustic methods. Together, these approaches provide complementary coverage across species and observation conditions.
We further outline how contextual environmental information such as water quality data informs adaptive sampling strategies to guide where and when observations are prioritised, improving efficiency without exhaustive survey coverage.
By integrating multi-modal sensing with adaptive operations, this work highlights how autonomous systems support scalable, data-driven marine species monitoring and effective ocean observations.