The beauty of remote sensing is that it can connect the world and “helps us understand our world a little better” as former Trimble eCognition Product Manager, Christian Weise once said.
Our customer ENEA, from Italy, is supporting an international project agreed upon between the Italian Ministry of the Environment and Land and Sea Protection (MATTM) and the Maritime and Ocean Affairs Division of the Ministry of Foreign Affairs of the Republic of Vanuatu (MAEV) by developing a Marine Spatial Plan through providing habitat mapping of the coastal marine environment through remote sensing.
ENEA’s contribution will focus on the automatic interpretation of Sentinel-2 satellite images to map the coastal marine environment of the Republic of Vanuatu in the Southern Pacific Ocean.
ENEA will generate an atlas of the Vanuatu archipelago on a scale between 1: 100,000 and 1: 50,000. The atlas will be based on freely available Sentinel-2 imagery analyzed with the Trimble eCognition software in combination with thematic data. The atlas will report the classification of marine habitats such as corals, seaweed and mangroves throughout the Republic. Mangroves are salt-tolerant evergreen plants and grow in the inter tidal coastal area of tropical and subtropical regions. Their worldwide distribution affects 118 countries, with a total global area of approximately 137,760 km2. These plants play an important role in the stabilization of the coast, river banks and in maintaining the ecological balance and biodiversity.
Mapping the extent of mangroves is important both for their protection and understanding their response to current climate change. However, accurately mapping the extent of this complex environment is still a challenge in the field of remote sensing. The Sentinel-2A satellite, launched in 2015, offers new opportunities for these research activities. The digital mapping of mangroves was performed partially using the procedure used by Wang et al. (2018), through the application of a two-level hierarchical structure using object-based image analysis (OBIA) within the eCognition Developer software.
The first level of the hierarchical classification approach allowed for the discrimination between vegetation from the non-vegetation, essentially the water, thus creating a vegetation mask. The second level then distinguished the mangrove cover within the vegetation mask.
The object-based classification essentially consisted of two phases: image segmentation and object classification. This last phase was obtained by developing a series of algorithms (rule sets) in which different spectral indices combined with each other were applied in order to obtain the best class description of this plant cover.