Robustness Verification of an Autonomous Underwater Vehicle-based Plankton Classifier

Jul 16, 2026

Upload by Hazel

The assessment of plankton standing stocks and microorganism structures is essential for understanding upper-ocean biological processes, yet in-situ optical classification remains unreliable because dynamic underwater conditions introduce noise and non-biological artifacts that trigger frequent misclassifications. This work presents a robustness verification framework for AUV-based plankton classifiers using reachability analysis, and a continuous-time neural ODE classification model built on high-resolution SilCam particle imaging. Formal verification establishes stability against environmental perturbations, enabling automated filtering of ambiguous data and reducing post-validation effort.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages6

Reading Time15 min

Introduction

Motivation and challenges in in-situ plankton imaging

Need for rigorous verification

Formal verification and reachability background

Neural Ordinary Differential Equations

Neural ODE model overview

Verification Framework

Robustness guarantees via reachability

Robustness Verification of an Autonomous Underwater Vehicle-based Plankton Classifier Abdelrahman Sayed Sayed Univ Gustave Eiffel, COSYS-ESTASF-59657 Villeneuve d’Ascq, France [abdelrahman.ibrahim@univ-eiffel.fr](abdelrahman.ibrahim@univ-eiffel.fr) Pierre-Jean Meyer Univ Gustave Eiffel, COSYS-ESTASF-59657 Villeneuve d’Ascq, France [pierre-jean.meyer@univ-eiffel.fr](pierre-jean.meyer@univ-eiffel.fr) arXiv :2607 .04453v 1 [ cs .RO] 5 Jul 2026 Asgeir J. Sørensen Department of Marine Technology Norwegian University of Science and Technology (NTNU) Trondheim, Norway [asgeir.sorensen@ntnu.no](asgeir.sorensen@ntnu.no) Mohamed Ghazel Univ Gustave Eiffel, COSYS-ESTASF-59657 Villeneuve d’Ascq, France [mohamed.ghazel@univ-eiffel.fr](mohamed.ghazel@univ-eiffel.fr) Abstract—The assessment of planktonic standing stocks and microorganism structures is critical for understanding upper ocean biological processes. Currently, autonomous underwater vehicles (AUVs) equipped with in-situ optical imaging and artificial intelligence (AI) methods offer a promising solution for persistent surveillance, mapping and monitoring of planktonic life. However, current AI methods often lack robustness in dynamic, unstructured environments, where environmental noise and non-biological artifacts lead to frequent misclassifications. Standard convolutional neural network (CNN) classifiers often struggle with such conditions, leading to misclassifications that require time-consuming manual validation by marine biologists. To address this issue, we propose a novel robustness verification framework for in-situ plankton classifiers based on reachability analysis. We also introduce a continuous-time neural ordinary differential equation (neural ODE) classification model leveraging the high-resolution imaging capabilities of the SilCamparticle imager. In this paper, we demonstrate the effectiveness of the proposed framework by formally verifying the robustness of the neural ODE model against environmental perturbations. We demonstrate that our verification framework acts as an automated filter providing formal guarantees of model stability against ambiguous data, thereby improving the reliability of autonomous sampling and reducing the post-processing workload. Index Terms—Neural ODE, CNN, Reachability analysis, Formal methods, Robustness verification, AUVs I. INTRODUCTION The growing interest in assessing and monitoring planktonic standing stocks and microorganism structures in the upper water column (mesopelagic zone) is critical for understanding the impact of climate change on ocean processes. To achieve continuous systematic ecosystem surveillance and monitoring, the oceanographic community is increasingly relying on autonomous underwater vehicles (AUVs) equipped This project has received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie COFUND grant agreement no. 101034248. This work is partly funded by SFI HARVEST, 309661 by the Research Council of Norway. with advanced optical sensors, such as the SilCam [1], and Artificial Intelligence (AI)-based tools. These AUVs enable high-resolution imaging and intelligent onboard sampling [2], significantly accelerating the detection and classification of microorganisms while simultaneously reducing the operational cost of deploying fleets of research vessels. Currently, the automated classification of in-situ plankton imagery is mostly dependent on convolutional neural networks (CNNs), including task-specific architectures such as ZooplanktoNet [3], and in-situ imaging and analysis pipelines built around the SilCam imager [1] and PyOPIA toolbox [4] . However, deploying AI-based classification models in uncertain and unstructured underwater environments raises significant challenges. Compared to controlled laboratory settings, in-situ ocean imaging is affected by dynamic environmental noise, turbidity, and particularly by gas bubble

FAQs

Why do current AI-based in-situ plankton classifiers misclassify images?

Dynamic environmental noise, turbidity, and bright non-biological artifacts such as gas bubbles from AUV motion often overlay the particles and lead to frequent misclassifications.

What modeling approach does the paper use for the classifier?

It introduces a continuous-time neural ordinary differential equation (neural ODE) model that leverages high-resolution imaging from the SilCam particle imager.

How does the robustness verification framework provide guarantees?

The framework combines neural ODE robustness properties with reachability analysis to formally verify stability of classification outputs within specified perturbation thresholds, acting as an automated filter for ambiguous data.

Jul 16, 2026

Upload by Hazel