In recent years, the demand for 3D spatial information and 3D city models has increased, as they support and allow many different applications, e.g. noise simulations, energy demand estimations, and shadow analysis. Constructing a city model with 3D buildings requires elevation data (such as LiDAR or Digital Terrain Models), but unfortunately, data of sufficient quality is often unavailable. This thesis focuses on the use of machine learning methods to estimate the height of building footprints and thus bypassing the use of elevation data completely. Three different methods are tested and compared: Random Forest Regression (RFR), Multiple Linear Regression (MLR), and Support Vector Regression (SVR). A case study is performed for the conterminous United States of America (USA) because of its availability of a nation-wide building dataset, containing roughly 125 million building footprints. The high diversity in urban layouts is considered, where a distinction is made between Central Business Districts (CBDs) in cities and all other regions (e.g. suburbs and rural areas). All building footprints are characterised by nine features derived from their geometry, which are then used (in several combinations) in the model training and predicting stages. Furthermore, the influence of additional features – including census and cadastral data – on the results of the building height predictions is analysed for the city of Denver, Colorado. The experiments show that it is feasible to predict the height for all buildings in the conterminous USA in under 6 minutes. Both the MLR and SVR method even accomplish it in under 30 seconds. The height prediction results show that the different prediction models struggle to accurately estimate the height for buildings in CBDs. The lowest achieved Mean Absolute Error (MAE) is 31.81m, whereas for the suburban and rural areas it is 1.41m. Adding additional, non-geometric features (e.g. census data) to the prediction models for one city (Denver) proved to be successful; the RFR method reduced its MAE from 1.35m to 0.96m for the suburbs, achieving sub-metre accuracy. The CBDs, however, are still problematic with an MAE of 16.87m. These results show that for the suburban and rural areas, the accuracy recommendations from the CityGML specifications for LOD1 models can be met (5m limit). For the CBDs, improvement is required. The experiments also proved that the proposed methodology can be used to generate 3D city models of very large datasets if no elevation data is available. Moreover, the method is, in theory, generic enough to be applied outside the USA.

The surface water extent around the world is constantly changing due to natural factors (e.g. geology and over-abstraction of water), climate change (e.g. higher water evaporation due to warmer climate) or human activities (e.g. reservoir construction). Water reservoirs are important for the management of the ecosystem, as both humans and the natural environment depend highly on them for their existence and well being. Flood control, agricultural irrigation, electricity generation, drinking and municipal water supply are only some of their main uses. Considering this, it is of high importance to have accurate maps that depict the reservoir outlines to determine their surface extend and storage capacity. However, their extend is not always well defined or there are discrepancies between various surface water datasets. This dissertation aims to provide an answer about which datasets match better as well as identifying the problematic areas by performing a quality control analysis. The main challenge of this thesis is that all available datasets have certain limitations regarding their coverage and quality. The waterbody delineation from satellite images is affected by the atmospheric conditions (e.g cloud obstructions) or topographic elements that create artifacts and influence the correct classification of water pixels. OpenStreetmap (OSM) on the other hand, has uncertain quality over locations, as the data is freely supplied by volunteers. Moreover, HydroLAKES which was created based, amongst others, on the Global Reservoir and Dam Dataset (GRanD), is still incomplete. In this thesis, an intercomparison of accuracy algorithm that can perform large scale analysis is created, by using the country of Angola as a use case, five datasets as input (Global Surface Water, Sentinel 2, OSM, HydroLAKES, GRaND) in both raster and vector format and the cloud processing platform of Google Earth engine. The identification of similarities or mismatches between the datasets is performed in terms of positional accuracy. Two quality measures have been considered for the pairwise comparison of features: percentage of overlap and Hausdorff distance. In addition the completeness of the datasets respectively to the total common water area of the water reservoir datasets is reviewed. The results of this research shows that large scale analysis for the comparison of accuracy between water reservoir datasets of different formats is possible. The pre-processing of the input Satellite data is semi-automated. The created automated algorithm for the main analysis offers information for all corresponding features between datasets. More specifically, statistics about the shape similarity, the percentage of overlap and the water area completeness of the datasets are being presented.