I. Introduction
Over the last 50 years, the world has faced dramatic growth of its urban population. In 2007 the amount of urban residents has outnumbered the rural population for the first time in history and by the year 2030 already two-thirds of the world’s population is expected to live in cities [1]. The number of so-called mega cities –cities with more than 10 million inhabitants– increased in the period from 1975 until today from 4 to 22, mostly in less developed regions [2]. Furthermore, a United Nations report [1] projects that almost all global population growth in the next 30 years will be concentrated in urban areas. This dynamic phenomenon will necessitate advanced methodologies such as space technologies, which will help city planners, economists, environmentalists, ecologists and resource managers solve the problems which accompany such growth [3]. In this uncontrolled situation, city planners lack tools to measure, monitor, and understand urban sprawl processes.
Multitemporal remote sensing has become an important data-gathering tool for analysing these changes [4]. In our study we use multi-sensoral time-series of Landsat MSS, Landsat TM and Landsat ETM data as well as TerraSAR-X stripmap data to continuously classify urban footprints from the 1970s to 2008. We use hierarchical, object-oriented classification methodologies for automatic detection of urbanized areas. This lets us detect temporal and spatial urban sprawl, densification processes and urban development on city level in our test sites: the European mega cities of Moscow, Russia and Istanbul, Turkey.
II. Data
III. Methodology
IV. Results
The analysis reveals dichotomies in spatiotemporal urban development for the two study sites. While the urban footprint of Istanbul is determined by the coastal and hilly orography, the urban footprint of Moscow is not subject to orographic restrictions. Differences include absolute urbanized area (Moscow by far larger), different temporal evolution (extensive urban sprawl in Moscow began post 2000, while in Istanbul extensive urban sprawl until 2000 has now been replaced by densification and flattening spatial gain), dispersed axial growth (Moscow) versus laminar growth (Istanbul). The comparison of complete urbanized area to official absolute population –Moscow 10.9 million and Istanbul 11.1 million – reveals different living characteristics for the two mega cities. In addition, we compared urban patterns using six concentric ring zones around the main urban centre (Figure 3). We use gradient analysis to identify location- and time dependent areal growth effects comparing the urban core to the periphery in the two mega cities. We see a clear shift of spatial growth to peripheral areas and more or less stagnant urban centres.
Due to a large amount of mixed spectral information in such a coarse ground resolution of the Landsat imageries the accuracy is limited. But for the requirement of mapping the city footprint, its spatial dimension and the spatial developments over the years, the Landsat images provide enough information for an assessment of urban change. An accuracy assessment has been performed by a randomization of 150 checkpoints and a visual verification process. The accuracy assessment for the various scenes resulted in 89 to 92 %. Using the presented approach urban footprints can be derived from TSX images with a total accuracy of 91 to 95%.
V. Conclusion
The study shows the capabilities of TerraSAR-X data to continue monitoring of urban sprawl processes to date. Although the capabilities of the two data types –Landsat and TerraSAR-X– vary in geometric and radiometric details, the individual classifications of urban footprints achieve comparable results for change detection. With classification accuracies around 90 % for the individual images the multi-sensoral data sets enable to assess the correct dimensions of urban growth, its directions and the large-area patterns. Especially in the highly dynamic urban environments the possibility of continuative monitoring of urbanization is essential for substantial decision-making.
Acknowledgements
References
link para o texto integral:
http://sss.terrasar-x.dlr.de/papers_sci_meet_3/poster/LAN002_taubenboeck.pdf
Sem comentários:
Enviar um comentário