Author(s): Silvia Huber; Lisbeth Tangaa Nielsen; Malinda Auluck; Muhammad Adzmin Bin Abdul Fatta; Lin Ji Liaw; Mikkel Bojesen; Lars Boye Hansen
Linked Author(s): Malinda Auluck
Keywords: Copernicus Sentinel-2 Deep learning Local knowledge Marine coastal habitats
Abstract: With increasing climate pressures and human impacts, the coverage and health of coastal ecosystems have declined globally. These ecosystems play an essential role for food provision and biodiversity and they act as Nature-based Solutions for erosion control and disaster resilience. Due to the essential ecosystem functions and to ensure sustainable management, accurate and up-to-date information on their distribution and condition is essential. However, such data is often lacking, and large-scale monitoring is challenging due to a missing link to local expertise. To address this gap, we developed a Human-In-the-Loop (HIL) cloud-based interactive platform enabling local experts to utilize their in-depth knowledge of their local marine coastal habitats in combination with open access satellite imagery and deep learning algorithms to create up-to-date classifications of Mangroves, Corals and Submerged Aquatic Vegetation in just a few minutes. We call this platform MCSAV. The MCSAV platform has been implemented for the Semporna reef complex in Malaysia, which is known for its rich marine life, to support the management and conservation of this critical aquatic biotope in response to development pressures, particularly tourism. However, the approach is scalable to any coastal environment.
Year: 2025