Article: Selection of the best hybrid spectral similarity measure for characterising marine oil spills from multi-platform hyperspectral datasets Journal: International Journal of Computational Science and Engineering (IJCSE) 2023 Vol.26 No.6 pp.715 - 731 Abstract: Marine oil pollution causes major economic crises in major industrial sectors like fishing, shipping and tourism. It affects marine life and human even decades after spillage necessitating very quick detection and remediation. Generally, oils are exceedingly difficult to identify from high-resolution images as oil slicks and sea water possess identical spectral characteristics. Therefore a cohesive and synergistic classification method called hybrid spectral similarity measures (HSSM) that discerns the data-rich constituents of hyperspectral images (HSI) is recommended in this paper to classify oil spills. Oil spill HSI procured from spaceborne [earth observation (EO-1) Hyperion] and airborne [airborne visible/infrared imaging spectrometer (AVIRIS)] platforms are employed to discriminate different marine spectral classes. The statistical parameters like overall accuracy (OA), Kappa, ROC/PR curve, AUC/PRAUC, weighted Youden index (J<SUB align="right"><SMALL>w</SMALL></SUB>), F1score and Noise analysis have identified spectral information divergence-chi square distance (SID-CHI) as the best HSSM promulgating its multi-class, multi-sensor, and multi-platform oil spill classification capability. Inderscience Publishers - linking academia, business and industry through research

Title: Selection of the best hybrid spectral similarity measure for characterising marine oil spills from multi-platform hyperspectral datasets

Authors: Deepthi; Deepa Sankar; Tessamma Thomas

Addresses: Department of Electronics, Cochin University of Science and Technology, Kochi, Kerala, 682022, India ' Division of Electronics Engineering, School of Engineering, Cochin University of Science and Technology, Kochi, Kerala, 682022, India ' Department of Electronics, Cochin University of Science and Technology, Kochi, Kerala, 682022, India

Abstract: Marine oil pollution causes major economic crises in major industrial sectors like fishing, shipping and tourism. It affects marine life and human even decades after spillage necessitating very quick detection and remediation. Generally, oils are exceedingly difficult to identify from high-resolution images as oil slicks and sea water possess identical spectral characteristics. Therefore a cohesive and synergistic classification method called hybrid spectral similarity measures (HSSM) that discerns the data-rich constituents of hyperspectral images (HSI) is recommended in this paper to classify oil spills. Oil spill HSI procured from spaceborne [earth observation (EO-1) Hyperion] and airborne [airborne visible/infrared imaging spectrometer (AVIRIS)] platforms are employed to discriminate different marine spectral classes. The statistical parameters like overall accuracy (OA), Kappa, ROC/PR curve, AUC/PRAUC, weighted Youden index (J_w), F1score and Noise analysis have identified spectral information divergence-chi square distance (SID-CHI) as the best HSSM promulgating its multi-class, multi-sensor, and multi-platform oil spill classification capability.

Keywords: hybrid spectral similarity measure; HSSM; hyperspectral image; HSI; ROC curve; weighted Youden index; F1 score; optimal cut-off value; OCV; signal to noise ratio; SNR; Hyperion; AVIRIS; oil spill.

DOI: 10.1504/IJCSE.2023.135285

International Journal of Computational Science and Engineering, 2023 Vol.26 No.6, pp.715 - 731

Received: 18 Dec 2021
Received in revised form: 09 Aug 2022
Accepted: 30 Aug 2022
Published online: 04 Dec 2023 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article

Title: Selection of the best hybrid spectral similarity measure for characterising marine oil spills from multi-platform hyperspectral datasets

Keep up-to-date