A GIS and object based image analysis approach to mapping the greenspace composition of domestic gardens in Leicester, UK

《A GIS and object based image analysis approach to mapping the greenspace composition of domestic gardens in Leicester, UK》

打印

作者

来源: LANDSCAPE AND URBAN PLANNING,Vol.183,P.133-146

语言: 英文

关键字: Domestic gardens; Object based image analysis; Urban greenspace; Worldview-2; LAND-COVER; ORIENTED CLASSIFICATION; ECOSYSTEM SERVICES; URBAN; VEGETATION; LANDSCAPE; ROADSIDE; IMPACTS; STORAGE; EXTENT

作者单位: [Baker, Fraser; Smith, Claire] Univ Leicester, Dept Geog, Univ Rd, Leicester LE1 7RH, Leics, England. [Baker, Fraser] Manchester Metropolitan Univ, Sch Sci & Environm, John Dalton Bldg,Chester St, Manchester M1 5GD, Lancs, England. Baker, F (reprint author), Univ Leicester, Dept Geog, Univ Rd, Leicester LE1 7RH, Leics, England.; Baker, F (reprint author), Manchester Metropolitan Univ, Sch Sci & Environm, John Dalton Bldg,Chester St, Manchester M1 5GD, Lancs, England. E-Mail: fraser.a.baker@stu.mmu.ac.uk; cls53@leicester.ac.uk

摘要: Greenspace provides a range of environmental benefits for urban residents, and is considered an important resource in urban development strategies. However, city-wide greenspace benefit modelling approaches often overestimate greenspace within garden areas, or even omit gardens from such analysis. As combined garden areal extents are significant for UK urban areas, improved estimations of garden greenspace abundance are required to improve urban greenspace analysis. This study investigates the methodological implications of a GIS and Object Based Image Analysis (OBIA) remote sensing approach with very high resolution Worldview-2 imagery for classifying UK urban garden surfaces. Gardens for an approximate 7 km(2) study area of the city of Leicester, UK were classified with an overall accuracy similar to 86%. The study demonstrates the applicability of GIS and OBIA analysis for mapping UK urban garden surfaces, with the methodology detailed here useful as a framework for further UK garden studies. Improvements to the current methodology are also considered in lieu of data and methodological limitations. In addition, the resulting garden surface dataset was analysed to examine associations of garden greenspace proportions with physical garden characteristics. Low greenspace proportions are found to be particularly prevalent for Victorian Terraced housing types in the study area; with greenspace proportions also generally associated positively with garden areal extents. Identification of potential predictive characteristics for garden greenspace abundance may prove useful as proxy information for urban greenspace analysis in other UK urban areas.