ABIA (2013 – 2016)

Agent Based Image Analysis

Object Based Image Analysis (OBIA) has meanwhile established well as a method for analyzing remote sensing data. However, the content of remote sensing data can be very complex, and so do the underlying classification rules. This hampers the transferability of rule sets and therefore the automation of OBIA when analyzing similar images: for each image the rule set needs to be adapted individually mirroring the imaging situation. To overcome this drawback we investigate the coupling, extension and integration of the agent-based paradigm with OBIA. (Software-)agents can act autonomously in complex environments and react flexible to unpredictable changes. Within a Multi Agent System (MAS) they can interact and cooperate with each other in order to achieve common and individual goals. These abilities allow agent-based systems to deal flexible and robust with complex and unpredictable situations. Against this background we hypothesize that we can increase the robustness and transferability of OBIA by integrating the agent-based paradigm with OBIA. Necessary rule-set adaptations and object-individual shape adaptations could be performed autonomously. In particular we will investigate: how agent-based principles can be transferred to OBIA and to what degree we need to extend recent OBIA methods, and: how can we realize agent-based principles in already existing OBIA environments.

Instrument: FWF, Stand-alone project

Role: Coordinator

Project volume: 275,000 EUR, ZGIS share: 225,000 EUR

Contact person: Thomas Blaschke, Peter Hofmann,

Researchers involved: Mariana Belgiu

Project website: www.zgis.at/abia