What question did this study set out to answer?

This research aims to develop a real-time environmental monitoring system for detecting hydrocarbon spills in marine infrastructures.

July 14, 20260 citationsOpen Access

View Full Paper

Real-Time Environmental Monitoring Dashboard and Multi-Modal AI Detection System for Marine Oil and Gas Facilities

NDNkue Dumka EEEmmanuel Iwok Essien UIUliang Ibia Ibia

Key Points

This research aims to develop a real-time environmental monitoring system for detecting hydrocarbon spills in marine infrastructures.
Implemented a multi-modal fusion system combining computer vision and Internet of Things (IoT) telemetry.
Used a YOLO deep learning network trained on high-resolution UAV imagery and SAR data.
Developed a React.js web dashboard for real-time visualization and alert generation.
Achieved an overall validation accuracy of 94.6% and an F1-score of 0.945.
Demonstrated a frame processing latency of 22 milliseconds.
Systemic end-to-end telemetry response time was 1.84 seconds, enhancing operational efficiency.

Abstract

Abstract: Marine oil and gas infrastructures are continuously exposed to the risk of hydrocarbon spills, necessitating rapid, automated surveillance frameworks to mitigate ecological catastrophes. This paper presents a multi-modal, automated real-time environmental monitoring system that fuses computer vision with localized Internet of a Things (IoT) biochemical telemetry. The detection architecture is powered by an optimized lightweight deep learning network (YOLO) trained on a heterogeneous dataset compiling high-resolution Unmanned Aerial Vehicle (UAV) RGB imagery and Synthetic Aperture Radar (SAR) Ground Range Detected (GRD) data from NASA Earthdata and the Copernicus Sentinel Data Hub. Simultaneously, an internet telemetry gateway continuously processes aquatic health parameters, including Water pH, Turbidity, and Dissolved Oxygen (DO). The backend infrastructure is built using the Django REST Framework to handle asynchronous image preprocessing, model inference, and a localized alert generation mechanism. Real-time visualization is rendered via a decoupled React.js single-page web dashboard. Empirically, the AI classification layer achieved an overall validation accuracy of 94.6% and an F1-score of 0.945, demonstrating a frame processing latency of 22 milliseconds. Combined with a systemic end-to-end telemetry response time of 1.84 seconds, the proposed framework offers an efficient, early-warning operational pipeline for environmental regulatory authorities and asset operators. Keywords: Oil Spill Detection, You Only Look Once (YOLO), Internet of a Things (IoT), Multi-Modal Fusion, Remote Sensing, Environmental Monitoring Dashboard. Title: Real-Time Environmental Monitoring Dashboard and Multi-Modal AI Detection System for Marine Oil and Gas Facilities Author: Nkue Dumka, Emmanuel Iwok Essien, Uliang Ibia Ibia, AniebietAbasi Prosper Daniel, Hameedat Bashir, Iwundu Elizabeth Chigozirim International Journal of Computer Science and Information Technology Research ISSN 2348-1196 (print), ISSN 2348-120X (online) Vol. 14, Issue 3, July 2026 - September 2026 Page No: 17-23 Research Publish Journals Website: www.researchpublish.com Published Date: 11-July-2026 DOI: https://doi.org/10.5281/zenodo.21319056 Paper Download Link (Source) https://www.researchpublish.com/papers/real-time-environmental-monitoring-dashboard-and-multi-modal-ai-detection-system-for-marine-oil-and-gas-facilities

Ask AI

Helpful

Bookmark

View Full Paper

Ask AI

Helpful

Bookmark

View Full Paper

Real-Time Environmental Monitoring Dashboard and Multi-Modal AI Detection System for Marine Oil and Gas Facilities

Key Points

Abstract

Cite This Study