Reconstructing urban scenarios for computational fluid dynamics simulations typically requires significant manual effort, especially when higher geometrical details are required. To address this issue, we present a workflow to automatically reconstruct buildings in three levels of detail (LoDs): LoD1.2, LoD1.3, and LoD2.2, tailored to urban microscale simulations. The workflow uses a combination of building footprints and a point cloud to segment roof planes, create partitions, optimise planes, and finally assemble roof planes into 3D building models. Reconstructed buildings are seamlessly integrated into the terrain together with different surface layers such as water, low vegetation, and paved surfaces. Apart from three general LoDs, building footprints can be simplified as a part of the 2D generalisation; additionally, smaller surfaces such as chimneys and ventilation shafts can be removed using a graph-cut optimisation. The integrated geometry validator can report on validity of building models, such as watertightness, manifoldness, or occurrences of self-intersections. In the case of invalid geometries, we can generate an approximation: geometry repair the with alpha wrapping algorithm, or reconstruction in lower LoD. We tested our implementation on two different real-world datasets — one in The Netherlands, and another one in the USA. The results showed that 95% (Dutch dataset) and 90% (US dataset) buildings were valid according to the ISO 19107 standard. Generated grids showed satisfactory quality as we observed monotonous convergence in simulations with grid convergence indices up to 3.8% for pressure and velocity variables. These results indicate that the workflow is suitable for typical urban microscale simulations.@en

We introduce CityJSON Text Sequences (CityJSONSeq in short), a format based on CityJSON and JSON Text Sequences. CityJSONSeq was added to the CityJSON specifications version 2.0 to allow us to stream very large 3D city models. The main idea is to decompose a CityJSON dataset into its individual city objects (each building, each tree, etc.) and create several independent JSON objects of a newly defined type: CityJSONFeature. We elaborate on the engineering decisions that were taken to develop CityJSONSeq, we present the open-source software we have developed to convert to and from CityJSONSeq, and we discuss different aspects of the new format, eg filesize, usability, memory footprint, etc. For several use-cases, we consider CityJSONSeq to be a better format than CityJSON because: (1) once serialised it is about 10% more compact; (2) it takes an order of magnitude less time to process; and (3) it uses significantly less memory.@en

Digital Elevation Models (DEMs) are a necessity for modelling many large-scale environmental processes. In this study, we investigate the potential of data from two spaceborne lidar altimetry missions, ICESat-2 and GEDI—with respect to their vertical accuracies and planimetric data collection patterns—as sources for rasterisation towards creating global DEMs. We validate the terrain measurements of both missions against airborne lidar datasets over three areas in the Netherlands, Switzerland, and New Zealand and differentiate them using land-cover classes. For our experiments, we use five years of ICESat-2 ATL03 data and four years of GEDI L2A data for a total of 252 million measurements. The datasets are filtered using parameter flags provided by the higher-level products ICESat-2 ATL08 and GEDI L3A. For all areas and land-cover classes combined, ICESat-2 achieves a bias of −0.11 m, an MAE of 0.43 m, and an RMSE of 0.93 m. From our experiments, we find that GEDI is less accurate, with a bias of 0.09 m, an MAE of 0.98 m, and an RMSE of 2.96 m. Measurements in open land-cover classes, such as “Cropland” and “Grassland”, result in the best accuracy for both missions. We also find that the slope of the terrain has a major influence on vertical accuracy, more so for GEDI than ICESat-2 because of its larger horizontal geolocation error. In contrast, we find little effect of either beam power or background solar radiation, nor do we find noticeable seasonal effects on accuracy. Furthermore, we investigate the spatial coverage of ICESat-2 and GEDI by deriving a DEM at different horizontal resolutions and latitudes. GEDI has higher spatial coverage than ICESat-2 at lower latitudes due to its beam pattern and lower inclination angle, and a derived DEM can achieve a resolution of 500 m. ICESat-2 only reaches a DEM resolution of 700 m at the equator, but it increases to almost 200 m at higher latitudes. When combined, a 500 m resolution lidar-based DEM can be achieved globally. Our results indicate that both ICESat-2 and GEDI enable accurate terrain measurements anywhere in the world. Especially in data-poor areas—such as the tropics—this has potential for new applications and insights.@en

Coastal elevation data are essential for a wide variety of applications, such as coastal management, flood modelling, and adaptation planning. Low-lying coastal areas (found below 10 m +Mean Sea Level (MSL)) are at risk of future extreme water levels, subsidence and changing extreme weather patterns. However, current freely available elevation datasets are not sufficiently accurate to model these risks. We present DeltaDTM, a global coastal Digital Terrain Model (DTM) available in the public domain, with a horizontal spatial resolution of 1 arcsecond (∼30 m) and a vertical mean absolute error (MAE) of 0.45 m overall. DeltaDTM corrects CopernicusDEM with spaceborne lidar from the ICESat-2 and GEDI missions. Specifically, we correct the elevation bias in CopernicusDEM, apply filters to remove non-terrain cells, and fill the gaps using interpolation. Notably, our classification approach produces more accurate results than regression methods recently used by others to correct DEMs, that achieve an overall MAE of 0.72 m at best. We conclude that DeltaDTM will be a valuable resource for coastal flood impact modelling and other applications.@en

When it comes to storing 3D city models in a database, the implementation of the CityGML data model can be quite demanding and often results in complicated schemas. As an example, 3DCityDB, a widely used solution, depends on a schema having 66 tables, mapping closely the CityGML architecture. In this paper, we propose an alternative (called ‘cjdb’) for storing CityGML models efficiently in PostgreSQL with a much simpler table structure and data model design (only 3 tables are necessary). This is achieved by storing the attributes and geometries of the objects directly in JSON. In the case of the geometries we thus adopt the Simple Feature paradigm and we use the structure of CityJSON. We compare our solution against 3DCityDB with large real-world 3D city models, and we find that cjdb has significantly lower demands in storage space (around a factor of 10), allows for faster import/export of data, and has a comparable data retrieval speed with some queries being faster and some slower. The accompanying software (importer and exporter) is available at https://github.com/cityjson/cjdb/ under a permissive open-source license.@en

This paper presents a new algorithm for filling holes in Level of Detail 2 (LoD2) building mesh models, addressing the challenges posed by geometric inaccuracies and topological errors. Unlike traditional methods that often alter the original geometric structure or impose stringent input requirements, our approach preserves the integrity of the original model while effectively managing a range of topological errors. The algorithm operates in three distinct phases: (1) pre-processing, which addresses topological errors and identifies pseudo-holes; (2) detecting and extracting complete border rings of holes; and (3) remeshing, aimed at reconstructing the complete geometric surface. Our method demonstrates superior performance compared to related work in filling holes in building mesh models, achieving both uniform local geometry around the holes and structural completeness. Comparative experiments with established methods demonstrate our algorithm’s effectiveness in delivering more complete and geometrically consistent hole-filling results, albeit with a slight trade-off in efficiency. The paper also identifies challenges in handling certain complex scenarios and outlines future directions for research, including the pursuit of a comprehensive repair goal for LoD2 models to achieve watertight 2-manifold models with correctly oriented normals. Our source code is available at https://github.com/tudelft3d/Automatic-Repair-of-LoD2-Building-Models.git@en

Understanding the UHI effect in any city requires high-resolution temperature data. This data is often difficult to obtain as cities usually have only a few ground sensors, leaving large data gaps. To fill these gaps, we compare Landsat-derived land surface temperature (LST) with air temperature (Tair) measurements from urban weather stations in the two largest cities in the Netherlands. Previous studies of this kind have often been limited due to a few main factors: low spatial resolution, limited weather station data and small sample sizes (Chung et al., 2020, Mutiibwa, 2015; Sheng 2017; Xiong, 2017; Yang, 2020). As a result, findings have been inconsistent, albeit mostly promising. Addressing these issues and adding to Burnett and Chen’s (2021) extensive comparison on a regional scale in Ontario, Canada, we present a reproducible, code-based approach focusing on cities. Using 149 Landsat scenes and data from 33 urban weather stations in the Netherlands (24 in Amsterdam, 9 in Rotterdam) between 2013-2022, 1700 comparison points across all European seasons are established. We find that there is a strong positive and significant linear relationship between LST and Tair across the dataset (r = .89). OLS regression results indicate 80% of the Tair variation can be explained by the LST, with Tair increasing by 0.62°C for every 1°C increase in LST. Analyses were repeated to account for seasonality, each station's local climate zone (Stewart and Oke, 2012) as well as mean absolute error and root mean square error to interrogate the discrepancy, all of which will be highlighted in the presentation. Overall, our evidence suggests that LST can indeed be a suitable proxy for Tair and could consequently form an additional decision-making layer to assist climate monitoring and urban planning in the Netherlands as well as similar climates.@en

We introduce a novel deep learning-based framework to interpret 3D urban scenes represented as textured meshes. Based on the observation that object boundaries typically align with the boundaries of planar regions, our framework achieves semantic segmentation in two steps: planarity-sensible over-segmentation followed by semantic classification. The over-segmentation step generates an initial set of mesh segments that capture the planar and non-planar regions of urban scenes. In the subsequent classification step, we construct a graph that encodes the geometric and photometric features of the segments in its nodes and the multi-scale contextual features in its edges. The final semantic segmentation is obtained by classifying the segments using a graph convolutional network. Experiments and comparisons on two semantic urban mesh benchmarks demonstrate that our approach outperforms the state-of-the-art methods in terms of boundary quality, mean IoU (intersection over union), and generalization ability. We also introduce several new metrics for evaluating mesh over-segmentation methods dedicated to semantic segmentation, and our proposed over-segmentation approach outperforms state-of-the-art methods on all metrics. Our source code is available at https://github.com/WeixiaoGao/PSSNet.

@en

We propose an enhancement module called depth discontinuity learning (DDL) for learning-based multi-view stereo (MVS) methods. Traditional methods are known for their accuracy but struggle with completeness. While recent learning-based methods have improved completeness at the cost of accuracy, our DDL approach aims to improve accuracy while retaining completeness in the reconstruction process. To achieve this, we introduce the joint estimation of depth and boundary maps, where the boundary maps are explicitly utilized for further refinement of the depth maps. We validate our idea by integrating it into an existing learning-based MVS pipeline where the reconstruction depends on high-quality depth map estimation. Extensive experiments on various datasets, namely DTU, ETH3D, “Tanks and Temples”, and BlendedMVS, show that our method improves reconstruction quality compared to our baseline, Patchmatchnet. Our ablation study demonstrates that incorporating the proposed DDL significantly reduces the depth map error, for instance, by more than 30% on the DTU dataset, and leads to improved depth map quality in both smooth and boundary regions. Additionally, our qualitative analysis has shown that the reconstructed point cloud exhibits enhanced quality without any significant compromise on completeness. Finally, the experiments reveal that our proposed model and strategies exhibit strong generalization capabilities across the various datasets.@en

Historical maps are increasingly used for studying how cities have evolved over time, and their applications are multiple: understanding past outbreaks, urban morphology, economy, etc. However, these maps are usually scans of older paper maps, and they are therefore restricted to two dimensions. We investigate in this paper how historical maps can be ‘augmented’ with the third dimension so that buildings have heights, volumes, and roof shapes. The resulting 3D city models, also known as digital twins, have several benefits in practice since it is known that some spatial analyses are only possible in 3D: visibility studies, wind flow analyses, population estimation, etc. At this moment, reconstructing historical models is (mostly) a manual and very time-consuming operation, and it is plagued by inaccuracies in the 2D maps. In this paper, we present a new methodology to reconstruct 3D buildings from historical maps, we developed it with the aim of automating the process as much as possible, and we discuss the engineering decisions we made when implementing it. Our methodology uses extra datasets for height extraction, reuses the 3D models of buildings that still exist, and infers other buildings with procedural modelling. We have implemented and tested our methodology with real-world historical maps of European cities for different times between 1700 and 2000.@en

While three-dimensional (3D) building models play an increasingly pivotal role in many real-world applications, obtaining a compact representation of buildings remains an open problem. In this paper, we present a novel framework for reconstructing compact, watertight, polygonal building models from point clouds. Our framework comprises three components: (a) a cell complex is generated via adaptive space partitioning that provides a polyhedral embedding as the candidate set; (b) an implicit field is learned by a deep neural network that facilitates building occupancy estimation; (c) a Markov random field is formulated to extract the outer surface of a building via combinatorial optimization. We evaluate and compare our method with state-of-the-art methods in generic reconstruction, model-based reconstruction, geometry simplification, and primitive assembly. Experiments on both synthetic and real-world point clouds have demonstrated that, with our neural-guided strategy, high-quality building models can be obtained with significant advantages in fidelity, compactness, and computational efficiency. Our method also shows robustness to noise and insufficient measurements, and it can directly generalize from synthetic scans to real-world measurements. The source code of this work is freely available at https://github.com/chenzhaiyu/points2poly.

@en

Roads are important for many urban planning applications, such as traffic modelling and delivery vehicle routing. At present, most available datasets represent roads only as centrelines. This is particularily true for OpenStreetMap which provides, among many features, road networks at worldwide coverage. Furthermore, most approaches for creating more detailed networks, such as carriageways or lanes, focus on doing so from sources that are not easy to acquire, such as satellite imagery or LiDAR scans. In this paper we present a methodology to create carriageways based on OpenStreetMap's centrelines and open access areal representations (i.e. polygons) to determine which roads should be represented as two individual carriageways. We applied our methodology in five areas across four different countries with different built environments. We analysed the outcome in a delivery routing problem to evaluate the validity of our results. Our results suggest that this method can be effectively applied to create carriageways anywhere in the world, as long as there is sufficient coverage by OpenStreetMap and an areal representation dataset of roads.

@en

In the computational fluid dynamics simulation workflow, the geometry preparation step is often regarded as a tedious, time-consuming task. Many practitioners consider it one of the main bottlenecks in the simulation process. The more complex the geometry, the longer the necessary work, meaning this issue is amplified for urban flow simulations that cover large areas with complex building geometries. To address the issue of geometry preparation, we propose a workflow for automatically reconstructing simulation-ready 3D city models. The workflow combines 2D geographical datasets (e.g., cadastral data, topographic datasets) and aerial point cloud-based elevation data to reconstruct terrain, buildings, and imprint surface layers like water, low vegetation, and roads. Imprinted surface layers serve as different roughness surfaces for modeling the atmospheric boundary layer. Furthermore, the workflow is capable of automatically defining the influence region and domain size according to best practice guidelines. The resulting geometry aims to be error-free: without gaps, self-intersections, and non-manifold edges. The workflow was implemented into an open-source framework using modern, robust, and state-of-the-art libraries with the intent to be used for further developments. Our approach limits the geometry generation step to the order of hours (including input data retrieval and preparation), producing geometries that can be directly used for computational grid generation without additional preparation. The reconstruction done by the algorithm can last from a few seconds to a few minutes, depending on the size of the input data. We obtained and prepared the input data for our verification study in about 2 hours, while the reconstruction process lasted 1 minute. The unstructured computational meshes we created in an automatic mesh generator show satisfactory quality indicators and the subsequent numerical simulation exhibits good convergence behavior with the grid convergence index of observed variables less than 5%@en

Data on the number of floors is required for several applications, for instance, energy demand estimation, population estimation, and flood response plans. Despite this, open data on the number of floors is very rare, even when a 3D city model is available. In practice, it is most often inferred with a geometric method: elevation data is used to estimate the height of a building, which is divided by an assumed storey height and rounded. However, as we demonstrate in this paper with a large dataset of residential buildings, this method is unreliable: <70% of the buildings have a correct estimate. We demonstrate that other attributes and characteristics of buildings can help us better predict the number of floors. We propose several indicators (e.g. construction year, cadastral attributes, building geometry, and neighbourhood census data), and we present a predictive model that was trained with 172,000 buildings in the Netherlands. Our model achieves an accuracy of 94.5% for residential buildings with five floors or less, which is an improvement of about 25% over the geometric approach. Above five floors, our model has only a slight improvement on the geometric approach (5%). The main culprit is the lack of training data for tall buildings, which is uncommon in the Netherlands.

@en

3D-representaties van onze leefomgeving zijn belangrijk in toepassingen die helpen bij de planning, de inrichting en het beheer van de openbare ruimte. Met 3D-modellen kunnen simulaties worden uitgevoerd voor bijvoorbeeld geluid, energie, luchtkwaliteit, windcomfort, zicht en wateroverlast. Om deze simulaties te bedienen met dezelfde 3D-gegevens, zijn standaarden essentieel. Voor bruikbaarheid in de praktijk moet een standaard voor 3D-data zowel specifiek als generiek zijn, eenvoudig gebruik van 3D-data mogelijk maken, en niet te grote bestanden opleveren. Dat is de motivatie geweest om CityJSON te ontwikkelen, een JSON-encoding (JavaScript Object Notation) van de OGC-standaard CityGML.@en

Recent developments in data acquisition technology allow us to collect 3D texture meshes quickly. Those can help us understand and analyse the urban environment, and as a consequence are useful for several applications like spatial analysis and urban planning. Semantic segmentation of texture meshes through deep learning methods can enhance this understanding, but it requires a lot of labelled data. The contributions of this work are three-fold: (1) a new benchmark dataset of semantic urban meshes, (2) a novel semi-automatic annotation framework, and (3) an annotation tool for 3D meshes. In particular, our dataset covers about 4 km² in Helsinki (Finland), with six classes, and we estimate that we save about 600 h of labelling work using our annotation framework, which includes initial segmentation and interactive refinement. We also compare the performance of several state-of-the-art 3D semantic segmentation methods on the new benchmark dataset. Other researchers can use our results to train their networks: the dataset is publicly available, and the annotation tool is released as open-source.

@en

Automated semantic segmentation and object detection are of great importance in geospatial data analysis. However, supervised machine learning systems such as convolutional neural networks require large corpora of annotated training data. Especially in the geospatial domain, such datasets are quite scarce. Within this paper, we aim to alleviate this issue by introducing a new annotated 3D dataset that is unique in three ways: i) The dataset consists of both an Unmanned Aerial Vehicle (UAV) laser scanning point cloud and a 3D textured mesh. ii) The point cloud features a mean point density of about 800 ​pts/m2 and the oblique imagery used for 3D mesh texturing realizes a ground sampling distance of about 2–3 ​cm. This enables the identification of fine-grained structures and represents the state of the art in UAV-based mapping. iii) Both data modalities will be published for a total of three epochs allowing applications such as change detection. The dataset depicts the village of Hessigheim (Germany), henceforth referred to as H3D - either represented as 3D point cloud H3D(PC) or 3D mesh H3D(Mesh). It is designed to promote research in the field of 3D data analysis on one hand and to evaluate and rank existing and emerging approaches for semantic segmentation of both data modalities on the other hand. Ultimately, we hope that H3D will become a widely used benchmark dataset in company with the well-established ISPRS Vaihingen 3D Semantic Labeling Challenge benchmark (V3D). The dataset can be downloaded from https://ifpwww.ifp.uni-stuttgart.de/benchmark/hessigheim/default.aspx.@en

Three-dimensional city models are essential to assess the impact that environmental factors will have on citizens, because they are the input to several simulation and prediction software. Examples of such environmental factors are noise (Stoter et al., 2008), wind (Garcı́a-Sánchez et al., 2014), air pollution (Ujang et al., 2013), and temperature (Hsieh et al., 2011; Lee et
al., 2013). However, those 3D models, which typically contain buildings and other man-made objects such as roads, overpasses, bridges, and trees, are in practice complex to obtain, and it is very time-consuming and tedious to reconstruct them manually. The software 3dfier addresses this issue by automating the 3D reconstruction process. It takes 2D geographical datasets (e.g., topographic datasets) that consist of polygons and “3dfies” them (as in “making them three-dimensional”). The elevation is obtained from an aerial point cloud dataset, and the semantics of the polygons is used to perform the lifting to the third dimension, so that it is realistic. The resulting 3D dataset is semantically decomposed/labelled based on the input polygons, and together they form one(many) surface(s) that aim(s) to be error-free: no self-intersections, no gaps, etc. Several output formats are supported (including
the international standards), and the 3D city models are optimised for use in different software.@en

As web applications become more popular, 3D city models would greatly benefit from a proper web-based solution to visualise and manage them. CityJSON was introduced as a JSON encoding of the CityGML data model and promises, among several benefits, the ability to be integrated with modern web technologies. In order to provide an implementation of a web application for CityJSON data, that can be used as a reference for other applications, we developed <code>ninja</code>. It is a web application that allows the user to easily load and investigate a CityJSON model through a web browser. In addition, it offers support for a complex feature of CityJSON: the experimental versioning mechanism. In this paper, we describe the motivation, requirements, technical aspects and achieved functionality of <code>ninja</code>. We believe that such a web application can facilitate the adoption of 3D city models by more practitioners and decision makers.

@en