Class Modelling Of Biotope Complexes–Success And Remaining Challenges

    Research output: Contribution to conferencePaperpeer-review

    Abstract

    Complex real world phenomena require means for reducing and modelling the underlying complexity. They ask for methods and tools to exploit the crucial information intelligently, for addressing dynamics and monitoring constant changes. The project Biotope Information and Management System (BIMS), carried out in a 3654 km² sized area in the region of Stuttgart/Germany, provides means for delineating and monitoring aggregated spatial units relevant for regional planning tasks. The project area is characterized by high dynamic growth and steady economic development. In the first phase of the project, we developed an adaptive per-parcel approach for delineating elementary units using SPOT-5 MS data (5 m GSD). In this paper we discuss our experiences from the second phase of the project, in which we developed algorithms for semi-automated class modelling of so-called biotope complexes based on these elementary units. The biotope complexes should represent highly aggregated, partly spectrally heterogeneous, but functionally homogenous units, following a mapping key provided by the Baden-Wuerttemberg Institute for Environmental Protection covering 19 different types of biotope complexes. Due to their very heterogeneous structure, some classes could not be successfully modelled and therefore not derived in a fully automated way. For that reason we developed a tool (biotope complex modeller -extension for ArcGIS) to support experts in aggregating the remaining units in a comfortable and fast manner. New research developments in the field of object-based accuracy assessment which emerged from this project will be discussed.
    Original languageEnglish
    Publication statusPublished - 2008
    EventGEOBIA 2008: Pixels, Objects, Intelligence GEOgraphic Object Based Image Analysis for the 21st Century - Calgary, Calgary, Canada
    Duration: 5 Aug 20088 Aug 2008
    https://www.isprs.org/proceedings/XXXVIII/4-C1/

    Conference

    ConferenceGEOBIA 2008
    Abbreviated titleGEOBAI
    Country/TerritoryCanada
    CityCalgary
    Period5/08/088/08/08
    Internet address

    Fields of Science and Technology Classification 2012

    • 105 Geosciences
    • 107 Other Natural Sciences

    Cite this