Increasing Chlorophyll-A Spatial Resolution Using Machine Learning
Вантажиться...
Дата
2024-10
Науковий керівник
Назва журналу
Номер ISSN
Назва тому
Видавець
IEEE
Анотація
This paper studied the question of measuring the concentration of chlorophyll-a in the Mediterranean Sea using satellite data. Low spatial resolution satellite data such as GCOM-C/SGLI and Sentinel-3 OLCI allow measurements of chlorophyll-a concentrations at the sea surface. However, these data have limited accuracy and spatial resolution, which creates challenges for monitoring local changes in coastal zones and small water areas. To increase the spatial resolution and accuracy of chlorophyll-a measurement, this paper proposes an information technology based on machine learning for a pilot area in the Mediterranean Sea near Cyprus. This technology combines low-resolution satellite data with ground-based measurements of chlorophyll-a and high-resolution data from the Sentinel-2 satellite. A comparative analysis of correlations between various satellite and ground data was carried out to determine the most informative features affecting the level of chlorophyll-a. Using the Random Forest and Multilayer Perceptron machine learning algorithms, an information technology was developed to improve the spatial resolution of chlorophyll-a concentration based on high-resolution satellite data. The developed technology makes it possible to create chlorophyll-a maps with a spatial resolution of 10 meters. The obtained results show a coefficient of determination of 0.36 and a correlation of 0.6 with ground measurements. The proposed approach is promising for monitoring the state of aquatic ecosystems.
Опис
Ключові слова
machine learning, satellite data, chlorophyll-a, cloud technologies, information technology, iMERMAID
Бібліографічний опис
Increasing Chlorophyll-A Spatial Resolution Using Machine Learning / Bohdan Yailymov, Pavlo Henitsoi, Nataliia Kussul, Andrii Shelestov // 2024 IEEE Fourth International Conference on System Analysis & Intelligent Computing (SAIC), [Kyiv], 8-10 October. - Kyiv, 2024. - 5 p.