Akademik Veri Yönetim Sistemi
2024 IEEE Mediterranean and Middle-East Geoscience and Remote Sensing Symposium, M2GARSS 2024, Virtual, Online, Cezayir, 15 - 17 Nisan 2024, ss.419-423
The marine mucilage outbreak of Spring 2021 in the inland Sea of Marmara raised significant environmental, economic and public-health related concerns. Recently, unmixing based approaches have been shown to provide spectral and spatial analysis of marine mucilage, in terms of endmembers and abundances. However, as the spectral signature of mucilage is prone to vary depending on factors such as its composition and density, and the characteristics of water in an inland sea are also highly dynamic, the linear mixture model (LMM) may be insufficient to correctly characterize the resulting mixture. This study proposes a Generalized Linear Mixing Model (GLMM) based unmixing approach to take spectral variability into account in unmixing based marine mucilage analysis. The proposed approach is evaluated on real PRISMA and DESIS hyperspectral data, and compared with LMM and Extended Linear Mixing Model (ELMM) based approaches.