Mapping of urban structures based on very high resolution satellite imagery for the city of DaNang, Vietnam

Kurzfassung

Detailled information on spatial structures is a key requirement for urban planning. However, many cities in developing and newly industrializing countries are lacking base data for their urban planning. Earth observation can therefore serve as a valuable source for administrations and urban planners. This work demonstrates the derivation of urban structures and building typologies for the city of DaNang, Vietnam, based on very high resolution (Pléiades) and high resolution (RapidEye) satellite imagery. Object-based image analysis was applied to identify built-up structures. Subsequently, data was combined with cadastral information to estimate the number of households in the city. Based on the "Local climate zones" scheme, nine different structures were classified at the building block level based on spatial parameters (such as size, density, shape or distances) and supervised machine learning classification. Derivation of surface models from the tri-stereoscopic Pléiades data by means of photogrammetric processing. Accurracy of both the classification and the surface stimates was assessed by using around 1000 reference points collected during field work. Concludingly, combining information of remote sensing, GIS data and geospatial analysis provides a thorough picture of the spatial structure of the city DaNang which can then be used for the planning of urban development.