Customer Stories

Classifying the Caribbean’s Coral

Issue link: https://geospatial.trimble.com/en/resources/i/1397539

Contents of this Issue

Navigation

Page 2 of 3

into eCognition. The software first delineates land and sea areas deeper than 15 m. Based on depth data and known topographic characteristics, it then categorizes the overall reef structure, distinguishing reef crest, fore reef, back reef, patch and fringing reef. Once it defines the five reef classes, eCognition determines seagrass classes (dense and sparse), sand types, dredged areas, and then finishes with mapping the deeper hard bottom algae classes. In total, the software automatically classifies 13 different benthic habitats. "eCognition is smart," said Ralph Humberg, Tama Group's managing director. "It takes in all this information we give it about coral reefs and uses it in the context of pixel characteristics and spatial relationships to classify features. It knows where features should and shouldn't exist. That's why it can accurately classify objects that look similar, like seagrass and hard bottom with algae." An aerial view of coral reefs off the southern coast of Haiti near Lagon des Huîtres National Park.. A Blue Hamlet swims in front of Staghorn Coral. Photo: Steve Schill Coral reefs surround Punta Icaco near Club Med Miches and Playa Esmeralda in the Dominican Republic. An eCognition classification of 12 different benthic habitats for Turks and Caicos.

Articles in this issue

view archives of Customer Stories - Classifying the Caribbean’s Coral