UDMV 2015
3rd Eurographics Workshop on Urban Data Modelling and Visualisation
November 23rd, 2015 · Delft, the Netherlands
3rd Eurographics Workshop on Urban Data Modelling and Visualisation
November 23rd, 2015 · Delft, the Netherlands
In many applications, such as in urban physical simulations or in the study of the effect of the solar impact at different scales, models with different levels of detail are required. In this paper we propose an efficient system for quickly computing the Sky View Factor (SVF) for any point inside a large city. To do that, we embed the city into a regular grid, and for each cell we select a subset of the geometry consisting of a square area centered on the cell and including it. Then, we remove the selected geometry from the city model and we project the rest onto a panoramic image (in our case, the sides of a box). Later, when several SVF evaluations are required, we only need to determine the cell that the evaluation point belongs to, and compute the SVF with the cell's geometry plus the environment map. To test our system, we perform several evaluations inside a cell's area, and compare the results with the ground truth SVF evaluation. Our results show the feasibility of the method and its advantages when used for a large set of computations. We show that our tool provides a way to handle the complexity of urban scale models, and specifically to study the sensitivity of the geometry.
Visibility analysis is an important application of 3D GIS data. Current approaches require 3D city models that are often derived from detailed aerial point clouds. We present an approach to visibility analysis that does not require a city model but works directly on the point cloud. Our approach is based on the medial axis transform, which models the urban environment as a union of balls, which we then use to construct a depthmap that is used for point visibility queries. As we demonstrate through our experiments on a real-world aerial LiDAR point cloud, the main benefits of our approach are 1) it is robust to noise, irregular sampling and holes of typical aerial LiDAR datasets, 2) it gives visibility results that are significantly more accurate than the often highly generalised city models, and 3) it is a simple algorithm that is easy to parallelise.
2D and 3D virtual architectural models are the common ground of many studies, including environmental protection, energy saving, or human well-being. Building or urban environment simulations concern for instance heat transfer, lighting, and acoustics, each of them requiring physical parameters additionally to the geometric representation. Furthermore, geometry does not generally comply straightforwardly with physical parameters and users are forced to manually adapt the models before simulation. This paper proposes an overview of modeling and simulation studies that make use of topological representations, and discusses the advantages of a topological representation for various types of applications. Such a representation can be used not only to maintain the 3D model global coherence, but also to automatically retrieve walls, doors, or room volumes for instance. Based on the existing model of generalized maps, this paper also illustrates some examples of structure traversal that can be used for providing the users with adequate simulation data.
The complexity of modern urban environments has led to the introduction of 3D Land Information Systems (LISs), which tend to replace traditional 2D LIS architectures for the purposes of urban planning and regeneration, land administration, real estate management and civil development. Both the need for 3D visualization of the geometry of buildings in various time instances through the years and the need for acquisition of 3D models in various levels of detail (LoDs), which not only fulfill the requirements of the various users but also they speed up the visualization process, are obvious. Thus, additional dimensions, that is, for time and scale, need to be supported by a modern LIS. This paper introduces a 5D modelling pipeline that may be adopted by a multi-purpose LIS for the selective creation of 3D models of an urban area in various time instances and at various LoDs, enriched with cadastral and other spatial data. The methodology is based on automatic change detection algorithms for spatial-temporal analysis of the spatial changes that took place in subsequent time periods, using image orientation, dense image matching and structure from motion algorithms; the procedure requires photogrammetric stereo plotting, implements procedural modelling and relies on the availability of overlapping aerial and terrestrial imagery, ground control points and cadastral information. Finally, an application based on the proposed methodology in an urban area in Greece is presented and the future work is discussed.
We investigate the automatic conversion between two substantially different formats used in 3D city models: the ubiquitous but semantically poor Wavefront OBJ and the semantically rich but less used OGC standard CityGML. We elaborate on their differences and on the challenges involved in their conversion, such as the inference of semantics in an OBJ file for their use in CityGML, and the storage of these semantics back in OBJ. We implement two software prototypes: a conversion of 3D building models from CityGML to OBJ (CityGML2OBJs), and one from OBJ to CityGML (OBJ2CityGML). By presenting both methods and implementations, we aim at increasing the availability of CityGML datasets and the possibility to create them in powerful 3D modelling software.
More than half of the world population lives in cities today. This proportion rises to 80% in developed countries. The density of urban population causes environmental troubles such as noise, urban heat waves, and chemical pollutions or magnetic pollution. Sensors and models are used to improve knowledge related to these phenomena particularly in cities. The aim of our research is to propose methods to view these phenomena in contextualised ways and at different levels of details. In the context of data exploration, we wish to generate from the initial phenomena data other levels of detail to allow the visual perception of the information at different visual scale. We also propose symbols that resist as well as possible to scale change and without excessive covering the other information such as streets, buildings or names. The first solutions presented in this paper are implemented and illustrated through two examples: nocturne temperature in Paris with very sparse initial data and concentration of chlorine with very dense initial data.
In this paper we describe a concept to manage and develop a web based virtual 3D scene, based on CityGML LoD 2 models, DTM tiles, ortho-photos and energy simulation results of specific heating demand and photovoltaic potential generated from SimStadt simulation platform, by integrating it on ESRI 3D City Information Model (3DCIM) platform. The final output results into a web based 3D visualization of multiple layers of building attributes such as building age, building height, building type, building usage and energy simulation results in terms of specific heating demand and PV potential. Additionally 3D modelling of trees and waterbody were produced based on its location to visually enrich the final virtual 3D scene.
Today, more and more cities worldwide are realizing the importance of semantic 3D city models. Various application areas of 3D city models such as simulations require the usage of highly dynamic and time-varying attributes, which are currently not supported by any standard. In this paper, we propose a new concept 'dynamizer', which extends static 3D city models by supporting variations of individual feature properties and associations over time. It allows to inject dynamic variations of city object properties into the static representation. In addition, the concept allows to model and study complex patterns representing dynamic variation of properties based on statistics and general rules.
Emotion assessments traditionally take place in academic or medical settings. The impact of environmental factors such as e.g. the characteristics of the place, the weather, or the time of day on the emotional state of inhabitants are often dismissed, thus emotions as a dynamic feature cannot be used for city planning or development. To estimate emotions as spatio-temporal data, a growing research community focused on emotion extraction from social network platforms such as Twitter or Facebook. The quality and the reliability of the resulting emotion evaluations are not comparable to direct emotion assessments, e.g. because of possible biases of emotional statements in social communities. In order to bridge this gap, we designed a web-based survey service that allows for the acquisition of geo-located emotion ratings of a huge group of participants in real-world environments, which in turn can be related to numerous environmental variables. In this paper, we present the architecture of this web-based service along with first results of a pilot study. In the end, we discuss possible use cases for spatio-temporal emotion data, the limitations of our web-based service, and present an outlook on possible future projects.
Activity recommendation systems aims at providing relevant information depending on targeted users' groups. For instance in a city, it makes sense to differentiate local residents from tourists. This research investigates to what extent the anonymized data collected from social networks can be used as a basis for making activity recommendations associated with local residents versus tourists when visiting a public place, such as a museum or gallery. Using rules based on the spatial, temporal and semantics of visited places, we are able to infer if a user is likely to be local or a tourist, based on anonymous sample Foursquare data and place-based semantics retrieved using Google Places API. Using semantics of visited places, it becomes possible to infer additional information about a user based on their movements over space and time. Depending on the kind and frequency of visited places, inferences about the aim of a visit to a location are possible. This analysis could provide information to users in the form of recommendations based on their movements while travelling around an area. This study has been performed using Foursquare check-ins for visitors to the Art Institute of Chicago between March 2010 and January 2011.
Simulating the evolution of urban landscapes is a challenging objective with a large impact not only for Computer Graphics (for its applications in the filming and gaming industries), but also for urban planning, economical and historical studies, urban physics, and many other. However, this target has remained elusive because of the large complexity implied by urban structures and their evolutions. We present a system that aims at simulating the evolution of the commercial structure in a modern city. In particular, given an initial distribution of shops, it studies the evolution when larger commercial areas, like malls, are introduced. This is computed using the Huff model as a measure of the attraction each commerce has on potential consumers, and an agent-based simulation to determine how these aspects affect their choices. Then, after a given simulation time, the system decides whether the shop has retained an income such that it can continue operating, or has gone bankrupt. Our system is used to study the evolution of the commercial structure of Barcelona city over the last century.