MULTI-TEMPORAL MULTI-SENSOR ANALYSIS OF URBANIZATION AND ENVIRONMENTAL/CLIMATE IMPACT IN CHINA FOR SUSTAINABLE URBAN DEVELOPMENT

摘要

The overall objective of this research is to investigaternmulti-temporal, multi-scale, multi-sensor satellite datarnfor analysis of urbanization and environmental/climaternimpact in China to support sustainable planning. Multitemporalrnmulti-scale SAR and optical data have beenrnevaluated for urban information extraction usingrninnovative methods and algorithms, including KTHPaviarnUrban Extractor, Pavia UEXT, and an “exclusioninclusion”rnframework for urban extent extraction, andrnKTH-SEG, a novel object-based classification methodrnfor detailed urban land cover mapping. Various pixelbasedrnand object-based change detection algorithmsrnwere also developed to extract urban changes. SeveralrnChinese cities including Beijing, Shanghai andrnGuangzhou are selected as study areas. Spatio-temporalrnurbanization patterns and environmental impact atrnregional, metropolitan and city core were evaluatedrnthrough ecosystem service, landscape metrics, spatialrnindices, and/or their combinations. The relationshiprnbetween land surface temperature and land-coverrnclasses was also analyzed.rnThe urban extraction results showed that urban areasrnand small towns could be well extracted usingrnmultitemporal SAR data with the KTH-Pavia UrbanrnExtractor and UEXT. The fusion of SAR data atrnmultiple scales from multiple sensors was proven tornimprove urban extraction. For urban land coverrnmapping, the results show that the fusion ofrnmultitemporal SAR and optical data could producerndetailed land cover maps with improved accuracy thanrnthat of SAR or optical data alone. Pixel-based andrnobject-based change detection algorithms developedrnwith the project were effective to extract urban changes.rnComparing the urban land cover results fromrnmulitemporal multisensor data, the environmentalrnimpact analysis indicates major losses for food supply,rnnoise reduction, runoff mitigation, waste treatment andrnglobal climate regulation services through landscapernstructural changes in terms of decreases in service area,rnedge contamination and fragmentation. In terms ofrnclimate impact, the results indicate that land surfacerntemperature can be related to land use/land coverrnclasses.