Abstract
Marine biodiversity monitoring requires scalability and reliability across complex underwater environments to support conservation and invasive-species management. Yet existing detection solutions often exhibit a pronounced deployment gap, with performance degrading sharply when transferred to new sites. This work establishes the foundational detection layer for a multi-year invasive species monitoring initiative targeting Arctic and Atlantic marine ecosystems.
We address this challenge by developing a Unified Information Pipeline that standardises heterogeneous datasets into a comparable information flow and evaluates a fixed, deployment-relevant detector under controlled cross-domain protocols. Across multiple domains, we find that structural factors — such as scene composition, object density, and contextual redundancy — explain cross-domain performance loss more strongly than visual degradation such as turbidity, with sparse scenes inducing a characteristic "Context Collapse" failure mode.
We further validate operational feasibility by benchmarking inference on low-cost edge hardware, showing that runtime optimisation enables practical sampling rates for remote monitoring. The results shift emphasis from image enhancement toward structure-aware reliability, providing a democratised tool for consistent marine ecosystem assessment.
Key Contributions
- Unified Information Pipeline: Standardises heterogeneous underwater datasets into a common, comparable information flow for cross-domain evaluation.
- Structural vs. Visual Degradation: Demonstrates that scene composition and object density predict cross-domain performance loss more strongly than image quality factors like turbidity.
- "Context Collapse" Failure Mode: Identifies and characterises a novel failure mode in sparse underwater scenes that disproportionately degrades detector performance.
- Edge Hardware Deployment: Benchmarks inference on low-cost edge devices, showing practical viability for remote marine monitoring at operational sampling rates.
BibTeX
@article{piccolo2026harmonizing,
title={Harmonizing the Deep: A Unified Information Pipeline for Robust Marine Biodiversity Assessment Across Heterogeneous Domains},
author={Piccolo, Marco and Han, Qiwei and van Toor, Astrid and Vanneste, Joachim},
journal={arXiv preprint arXiv:2601.13975},
year={2026},
url={https://arxiv.org/abs/2601.13975}
}