ROSID: Remote Sensing Satellite Data for Oil Spill Detection on Land

Nurseitov D.B. Abdimanap G. Abdallah A. Sagatdinova G. Balakay L. Dedova T. Rametov N. Alimova A.
December 2024 Engineered Science Publisher

Engineered Science
2024 #32

Oil spills on land pose significant environmental hazards, impacting ecosystems and human health. Effective detection and monitoring of these spills are critical for timely response efforts. This paper presents for the first time a ground truth dataset for onshore oil spill detection developed using Landsat imagery. The ground truth was implemented using aerial data. To demonstrate the utility of this dataset, we evaluated several state-of-the-art deep learning models, including DeepLabV3+, UNet, PSP-Net, DeepLabV3 and Mask2Former. Our experiments revealed significant insights into the models’ capabilities and limitations. Mask2Former and DeepLabV3+, in particular, showed the highest performance metrics. On the validation data, Mask2Former achieved an intersection over Union (IoU) of 72.69% and a F-score coefficient of 84.18%, while DeepLabV3+ achieved an IoU of 67.6% and a F-score coefficient of 80.67%. These results demonstrate the effectiveness of our dataset as a crucial tool for enhancing oil spill detection methodologies and advancing the application of artificial intelligence in ecological preservation and disaster management.

Deep Learning , Environmental monitoring , Image segmentation , Landsat imagery , Oil pollution dataset , Remote sensing , Terrestrial oil spill

Text of the article Перейти на текст статьи

KazMunayGas Engineering LLP, Astana, 010000, Kazakhstan
Institute of Ionosphere LLP, Almaty, 050020, Kazakhstan
Software Engineering Department, Satbayev University, Almaty, 050013, Kazakhstan
Department of Computer Science & DiSC, University of Innsbruck, 6020, Austria
Information Technology Department, Assiut University, Assiut, 71515, Egypt
Department of Geospatial Engineering, Satbayev University, Almaty, 050013, Kazakhstan

KazMunayGas Engineering LLP
Institute of Ionosphere LLP
Software Engineering Department
Department of Computer Science & DiSC
Information Technology Department
Department of Geospatial Engineering

10 лет помогаем публиковать статьи Международный издатель

Книга Публикация научной статьи Волощук 2026 Book Publication of a scientific article 2026