Classification and quantification of green in the expanding urban and semi-urban complex: Application of detailed field data and IKONOS-imagery

Urban land cover is expanding rapidly worldwide. This major phenomenon is often accompanied by an expansion of a green component. Urban green can itself be considered as a most important but often ignored land cover category. With this study we investigate how IKONOS data can be used more exhaustively for the detection and more importantly the quantification of urban green, compared to state-of-the art investigations. This paper demonstrates how a combination of specific techniques, including pansharpening, the use of vegetation indices and object detection can enhance the possibilities to map vegetated elements and even estimate volumes of woody patches in the southern fringe of Roeselare (Belgium). The values of the soil adjusted MSAVI index are found to be related to the increase in volume of the trees (coniferous and deciduous). To analyze the vegetation in more detail, we use an object-oriented classification with MSAVI to exclude the sealed areas from the further analysis. With a rule set of segmentation and classification steps, the vegetation is defined on a higher level. Especially textural measures are of importance to separate grass from high vegetation.