Extracting urban forests in GEOBIA framework using AGROVOC and WorldView-2 imagery

In GEOBIA, segmentation is a very first task in creating image objects from very high spatial resolution (VHSR) imagery. In general, the extracted image objets are readily be used in the GIS - ready vector format. However further investigations, developments and testing of methods in extracting accurate image objects are needed. In this paper, we have developed thematic and spatial semantics using AGROVOC Ontologies for extracting vocabulary for the agricultural domain, which is owned and maintained by an international community of Agricultural Research Information Institutions and Food and Agriculture Organization (FAO) of the United Nations. This has been published as Linked Open Data so that information can be used automatically. In this case study example, we aim to develop a proof-of-concept for classes like tree, tree types, plantation and forest in multiple scales. The preliminary results showed that in GEOBIA framework, objects can be characterised with semantic meaning and their relationship with the real world. This was tested for a Worldview-2 imagery of Hobart, Tasmania, Australia in depicting urban trees and other agricultural objects.