期刊名称：ISPRS Annals of the Photogrammetry, Remote Sensing and Spatial Information Sciences
印刷版ISSN：2194-9042
电子版ISSN：2194-9050
出版年度：2005
卷号：XXXVI-8/W27
出版社：Copernicus Publications
摘要：Satellite remote-sensing data for urban land -cover and land-use mapping has been investigated over a long time using increasingly more powerful data and more sophisticated techniques. Well-established methods for surveying urban habitats such as aerial photography have to be replaced now for updating previous results. At present additional benefits result from high resolution airborne scanner and satellite data. In this study primarily HRSC-AX and satellite data (QuickBird, IKON OS) are compared. Both visual and automatic interpretation using ERDAS Imagine and eCognition were tested. Study areas offering particular challenges concerning urban green and classification of sealing have been selected in Berlin and Seoul, representing different starting positions according to existing data, mapping methods and urban structures. For the purpose of updating the urban habitats in the study area Berlin the HRSC -AX imagery were classified by a combined approach of visual and automatic pixel-based classification. In Seoul, QuickBird and IKONOS images were classified by a stepwise pixel and segment-based method. As a result of these two case studies potentials of new remote sensing techniques and sensors for mapping urban habitats could be show n. Advantages and disadvantages e.g. colour shift, shadows or inclined recording of the available new data sets are discussed. Over all the visual interpretation of both aerial scanner and satellite data was best suited for habitat type mapping in the two study areas
关键词：Habitat M apping; High Resolution; HRSC -AX; IKONOS; QuickBird
