TY - JOUR A1 - Wurm, Michael A1 - Stark, Thomas A1 - Zhu, Xiao Xiang A1 - Weigand, Matthias A1 - Taubenböck, Hannes T1 - Semantic segmentation of slums in satellite images using transfer learning on fully convolutional neural networks JF - ISPRS Journal of Photogrammetry and Remote Sensing N2 - Unprecedented urbanization in particular in countries of the global south result in informal urban development processes, especially in mega cities. With an estimated 1 billion slum dwellers globally, the United Nations have made the fight against poverty the number one sustainable development goal. To provide better infrastructure and thus a better life to slum dwellers, detailed information on the spatial location and size of slums is of crucial importance. In the past, remote sensing has proven to be an extremely valuable and effective tool for mapping slums. The nature of used mapping approaches by machine learning, however, made it necessary to invest a lot of effort in training the models. Recent advances in deep learning allow for transferring trained fully convolutional networks (FCN) from one data set to another. Thus, in our study we aim at analyzing transfer learning capabilities of FCNs to slum mapping in various satellite images. A model trained on very high resolution optical satellite imagery from QuickBird is transferred to Sentinel-2 and TerraSAR-X data. While free-of-charge Sentinel-2 data is widely available, its comparably lower resolution makes slum mapping a challenging task. TerraSAR-X data on the other hand, has a higher resolution and is considered a powerful data source for intra-urban structure analysis. Due to the different image characteristics of SAR compared to optical data, however, transferring the model could not improve the performance of semantic segmentation but we observe very high accuracies for mapped slums in the optical data: QuickBird image obtains 86–88% (positive prediction value and sensitivity) and a significant increase for Sentinel-2 applying transfer learning can be observed (from 38 to 55% and from 79 to 85% for PPV and sensitivity, respectively). Using transfer learning proofs extremely valuable in retrieving information on small-scaled urban structures such as slum patches even in satellite images of decametric resolution. KW - slums KW - FCN KW - convolutional neural networks KW - deep learning KW - transfer learning Y1 - 2019 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-233799 VL - 150 ER - TY - JOUR A1 - Taubenböck, H. A1 - Weigand, M. A1 - Esch, T. A1 - Staab, J. A1 - Wurm, M. A1 - Mast, J. A1 - Dech, S. T1 - A new ranking of the world's largest cities—Do administrative units obscure morphological realities? JF - Remote Sensing of Environment N2 - With 37 million inhabitants, Tokyo is the world's largest city in UN statistics. With this work we call this ranking into question. Usually, global city rankings are based on nationally collected population figures, which rely on administrative units. Sprawling urban growth, however, leads to morphological city extents that may surpass conventional administrative units. In order to detect spatial discrepancies between the physical and the administrative city, we present a methodology for delimiting Morphological Urban Areas (MUAs). We understand MUAs as a territorially contiguous settlement area that can be distinguished from low-density peripheral and rural hinterlands. We design a settlement index composed of three indicators (settlement area, settlement area proportion and density within the settlements) describing a gradient of built-up density from the urban center to the periphery applying a sectoral monocentric city model. We assume that the urban-rural transition can be defined along this gradient. With it, we re-territorialize the conventional administrative units. Our data basis are recent mapping products derived from multi-sensoral Earth observation (EO) data – namely the Global Urban Footprint (GUF) and the GUF Density (GUF-DenS) – providing globally consistent knowledge about settlement locations and densities. For the re-territorialized MUAs we calculate population numbers using WorldPop data. Overall, we cover the 1692 cities with >300,000 inhabitants on our planet. In our results we compare the consistently re-territorialized MUAs and the administrative units as well as their related population figures. We find the MUA in the Pearl River Delta the largest morphologically contiguous urban agglomeration in the world with a calculated population of 42.6 million. Tokyo, in this new list ranked number 2, loses its top position. In rank-size distributions we present the resulting deviations from previous city rankings. Although many MUAs outperform administrative units by area, we find that, contrary to what we assumed, in most cases MUAs are considerably smaller than administrative units. Only in Europe we find MUAs largely outweighing administrative units in extent. KW - city size KW - urban agglomeration KW - rank-size distribution KW - remote sensing KW - global urban footprint KW - urban morphology Y1 - 2019 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-240634 VL - 232 ER - TY - JOUR A1 - Kaspar-Ott, Irena A1 - Hertig, Elke A1 - Kaspar, Severin A1 - Pollinger, Felix A1 - Ring, Christoph A1 - Paeth, Heiko A1 - Jacobeit, Jucundus T1 - Weights for general circulation models from CMIP3/CMIP5 in a statistical downscaling framework and the impact on future Mediterranean precipitation JF - The International Journal of Climatology N2 - This study investigates the projected precipitation changes of the 21st century in the Mediterranean area with a model ensemble of all available CMIP3 and CMIP5 data based on four different scenarios. The large spread of simulated precipitation change signals underlines the need of an evaluation of the individual general circulation models in order to give higher weights to better and lower weights to worse performing models. The models' spread comprises part of the internal climate variability, but is also due to the differing skills of the circulation models. The uncertainty resulting from the latter is the aim of our weighting approach. Each weight is based on the skill to simulate key predictor variables in context of large and medium scale atmospheric circulation patterns within a statistical downscaling framework for the Mediterranean precipitation. Therefore, geopotential heights, sea level pressure, atmospheric layer thickness, horizontal wind components and humidity data at several atmospheric levels are considered. The novelty of this metric consists in avoiding the use of the precipitation data by itself for the weighting process, as state-of-the-art models still have major deficits in simulating precipitation. The application of the weights on the downscaled precipitation changes leads to more reliable and precise change signals in some Mediterranean sub-regions and seasons. The model weights differ between sub-regions and seasons, however, a clear sequence from better to worse models for the representation of precipitation in the Mediterranean area becomes apparent. KW - CMIP3 KW - CMIP5 KW - Mediterranean area KW - precipitation KW - statistical downscaling KW - weights Y1 - 2019 U6 - http://nbn-resolving.de/urn/resolver.pl?urn:nbn:de:bvb:20-opus-325628 VL - 39 ER -