ASSESSMENT OF MACHINE LEARNING METHODS FOR SEAGRASS CLASSIFICATION IN THE MEDITERRANEAN

Gümüşay M. Ü., Bakırman T.

International Symposium on Applied Geoinformatics (ISAG-2019), İstanbul, Turkey, 7 - 09 November 2019, pp.485-488

  • Publication Type: Conference Paper / Full Text
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.485-488
  • Yıldız Technical University Affiliated: Yes

Abstract

ABSTRACT: Seagrasses are essential plants in marine ecosystems in regard to physical, chemical and biological cycles. They provide key functions for land and sea by producing and exporting organic carbon, regulating carbon dioxide, nutrient cycling, sediment stabilisation, preventing coastal erosion and reducing exposure to the bacterial pathogens of humans, fish and invertebrates. However, human activities such as mining, pollution and over fishing create a pressure on these plants and reduce the benthic biodiversity. Posidonia oceanica is an endemic seagrass species in the Mediterranean. Even though this species has been put under protection by EU legislation, the Bern and Barcelona onventions, and national legislation, according to the International Union for Conservation of Nature (IUCN) Red List of Threatened Species, P. oceanica will qualify as vulnerable no further recovery plans are established, especially in the Western Mediterranean region. Therefore, in order to conserve seagrasses high resolution, accurate and temporal distribution maps are need to be produced. In this study, it is aimed to create seagrass distribution maps with machine learning algorithms namely as random forests and support vector machines using WorldView-2 imagery. In-situ data has been collected via underwater video and scuba diving for classification training and testing. Atmospheric, radiometric and water column corrections are applied for pre-processing of optical satellite image. The light penetration in the water is limited by depth. Therefore, we have limited our study area based on maximum depth of 20 meters. The classification accuracies are calculated as 94% and 71% for random forests and support vector machines, respectively. According to the results, it can be clearly said that random forests method is superior to support vector machines for seagrass mapping in our study area. The proposed framework in this study enables to rapidly produce seagrass distribution maps which can be used monitor temporal change for sustainable ecosystem.

Keywords: Seagrass, Classification, Machine learning, Posidonia oceanica, Mediterranean