· · · Institutional Repository

With laser scanning (including laser altimetry and multi-beam echo sounding), many data points, called point clouds, are measured. The interest in point clouds is increasing. Depending on the laser scanner, the measurement set-up and the technique, the number of points in one dataset can vary from a few hundred points to over a billion points. The data volume of a laser survey mission can easily reach Gigabyte or even Terabyte level. It is a problem to handle these voluminous datasets efficiently. One way to manage the data is to partition the data with tiles or grids and then store the data in each tile or grid as a single file in text or binary format; the large volume of data is divided into separate files of a reasonable size. As an alternative, the data can be stored in a database management system (DBMS); each point in a single record based on standard data types. Since the data stored can be accessed in single point level, it is easy to perform queries and analysis on the data server, and only the qualified points are returned to the user. Since the data is spatial data, it has a location; data can be stored based on a spatial data type in a spatial DBMS, spatial objects can be clustered and spatially indexed to limit searches, with the result that spatial queries can be performed. Storing and retrieving massive datasets as single point records (one point per record) results in storing and retrieving lots of records, but fast retrieval become questionable, as there is overhead per record and because of the number of records to be retrieved. As Oracle is the market leader in the development of DBMS, this research is limited primarily to Oracle. This leads directly to the research question of this thesis: What is the best design for a data model to store large point clouds in an Oracle DBMS, such that it is generally accessible by spatial applications, that all attributes are preserved and that performance is optimised? Analysis A solution has been found by reducing the number of records by clustering nearby points without loss of information. If points are clustered, associated attributes have to be clustered as well. It has been decided how to record these attributes with the points. The Region Quadtree is implemented to grid the data in logical parts. The Morton space-filling curve is implemented to cluster the data. Hilbert space-filling curve would even be better. The points have been stored using a well-known data type called SDO_GEOMETRY (GTYPE 3005, POINTCLUSTER), SDO_GEOMETRY is a standard spatial data type implemented in Oracle. Attributes have been stored in one or more arrays (VARRAY), because ordinates are also stored in a VARRAY. This method is been compared to the newer SDO_PC data type. Conclusion The presented method using SDO_GEOMETRY (GTYPE 3005, POINTCLUSTER) performs optimal for large point cloud datasets in terms of performance and the points can generally be accessed by spatial applications. The alternative method using SDO_PC is not suitable for the conditions as stipulated in the research question. Because the SDO_PC data type is newly implemented in Oracle, not generally readable by spatial applications, not (yet) fully supported by all vendors and required storage space is double the amount.

This research discusses the conversion of underground utilities datasets into scene graph models for its 3D visualization using the WebGL technology, which is a Javascript interface to access the 3D graphics hardware. This conversion requires to reconstruct the volumetric appearance of the objects from its abstract representation, involving several decisions towards its mapping into a scene graph. Each decision taken carries consequences in the rendering performance and given the nature of the underground utilities, multiple problems arise which should be addressed before achieving a fluid visual- ization. This research considers multiple reconstruction approaches and studies the corresponding factors affecting the performance by introducing a testing framework. With the results of the tests, the considered factors are ranked based on their performance and used to give advice towards the best approaches for the reconstruction underground utilities and its applicability to WebGL environments.

With the rapid improvement of LIDAR systems regarding point density and accuracy in relation to the (application dependent) requirements, robustness, efficiency and automation of the modeling process are becoming more important than achieving the highest possible accuracy and modeling detail from the available LIDAR data. Therefore we opt for development of a 2D grid based LIDAR data analysis approach. An important step is detection and parameterization of planar surfaces (roof elements). The paper reviews four methods, based on analysis of gradients, principal components, least squares and hough transforms, respectively. It introduces a series of improvements to the standard usage of each of those methods and shows results from synthetic and real data.

The notion of terroir covers the interaction between the natural, the cultural conditions and the grape vine itself relative to the character of the final wine product. It is about a dynamic chain whose factors get different weights in each wine region and this results in the typicality of the different terroir wine products. The Old World wine regions have been for many years occupied with the terroir research in order to explain the uniqueness and the specificities of their wine products. Nowadays, more and more wine regions are interested in such kind of investigations in order to highlight their distinct products, expose their originality and get a place in the wine market. DOQ Priorat is located in South Catalonia in Spain and is a wine region that has recently been nominated a label of originality for its products. Such a denomination made the region’s cultivators interested to learn more about the natural conditions of the area in a try to explain the success of their products and with the prospect to preserve the quality of their wines. In these terms, Geographical Information Systems (GIS), providing the ability for spatial, statistical analysis and visualisation, have been considered important tools for the DOQ Priorat’s natural conditions’ analysis and diverse terroir investigation. Previous research has indicated topography, the soil properties and the climate of a region to define the natural conditions of a region and therefore being the factors of interest/effect in the wine terroir. These three factors can be described by specific attributes. For instance, elevation and ground inclination for topography, PH and texture for soil, temperature and precipitation for climate. The distribution of such attributes and many others that characterize the aforementioned natural conditions, has been studied in the DOQ Priorat territory in order to examine the resemblance of the DOQ Priorat vine growing conditions to the conditions that are considered beneficial from international research. The correspondence of the DOQ Priorat conditions to the standards for vine growing has been treated by means of a multivariate GIS analysis. The DOQ Priorat topographic and soil attributes have been evaluated relative to their suitability for vine growing and different suitability classes have been defined. Moreover, the growing season temperature distribution and its capability to define conditions of viability for specific grape varieties, has been used to define two major terroir units in DOQ Priorat. The suitability of the DOQ Priorat land for cultivation of specific varieties has been assessed through the topographic-soil composite suitability in relation to the teroir units defined. DOQ Priorat’s greatest extend has been classified in intermediate and intermediate to high vine growing suitability classes relative to its topography and soil conditions. The shallow and dry soils as well as the steep slopes visited in the area seem to contradict to what is considered beneficial for vine growing. In terms of climate however, the whole area presents ideal conditions for a wide variety of grape cultivations whereas most of these varieties proposed to fit the area are already cultivated today. Even though the current vineyard cultivations still present conditions of intermediate and intermediate-high suitability relative to what is considered beneficial from international vine growing research, the quality of the wines produced in DOQ Priorat is indisputable; there are therefore some unique features in the DOQ Priorat terroirs. The DOQ Priorat vines are cultivated in higher elevations and in soils shallower and less fine textured than what is considered to fit for vine growing internationally. That is what gives the DOQ Priorat wines their unique character and these are finally the conditions that are considered ideal for Priorat wines. Such a conclusion, leads to the confirmation of the dynamic nature of terroir, whose factors and attributes cannot be strictly defined and quantified for every wine region. New vine growing standards could be defined for several wine regions relative to their specificities whereas it is mostly the try, the result and the experience that define good and bad terroirs. The DOQ Priorat case has been a very nice example of a region going against the vine cultivation pattern whereas obtaining high quality and recognised wine products.

Nowadays the field of geo-information is undergoing major changes, and the transition from 2D to 3D is having a major influence. A significant amount of 3D datasets are stored in the database. Experts are aware that new quality control mechanisms need to be built into the database systems in order to secure and guarantee high-performing data. Constraints are effective in providing solutions needed to avoid errors and enable maintenance of data quality. Whereas constraints for 2D geographic datasets have already been the subject of several research projects, studies into 3D geo-data constraints are largely unexplored. This thesis research discovers a new approach to model, conceptualise and implement 3D geo-constraints which can function in the database. At the outset, constraints can be formulated using natural language. As natural language is subjective and varies between individuals, expressions can be ambiguous and can easily cause confusion. So spatial constraints are abstracted using geometry that depicts the exact shape, and also topology that reveals the spatial relationship between geometries. This step makes the meaning of a constraint clearer to others. Furthermore, using standardised UML diagrams and OCL expressions, geo-constraints can be formalised to an extent that not only humans, but also machines can understand them. With model-driven architecture supported by various softwares, OCL expressions can be automatically converted to other models/executable codes (e.g. PL/SQL) just by a few clicks. And with small modifications, database triggers can be formulated to carry out constraints check. A database including various topographic objects (e.g. buildings, trees, roads, grass, water-bodies and terrains) is used as a study case to apply the discovered approach. During this research, a first attempt to formulate 3D geo-constraints in OCL has been made. These expressions can be tested and translated to other models/implementations when the OCL standard is extended with spatial types and operations. In the implementation stage, the current 3D functions in Oracle Spatial database are found to be insufficient. A new 3D function using existing 2D functions - plus additional code relating to computational geometry - has been developed by the author to bridge the gap. Based upon this function, a large group of spatial constraints which apply to objects in 3D space can be checked. Bentley Map and Python IDLE are used to test the performance of constraints as well as the visualisation of warning messages to clients. Database error messages are immediately displayed on the front-ends when a modification that does not satisfy a constraint is attempted to commit to the database. During the case study, new classes of constraints are also discovered. They are higher-level constraints, parameterised constraints, constraints allowing exceptional instances, extra-check rules to detect conflicting constraints and constraints relating to multi-scale representations.

Knowledge of temperature distributions on streams and lakes is considered to be a valuable source of information for a wide range of disciplines such as ecologists, hydrologists and geochemists, as it can provide insights into the dynamics of these water bodies (Westhoff, 2006). Modeling of surface water temperature on the other hand is a complex process requiring coupling of spatial and hydrological data (Boyd, Kasper, 2003). At local scales, all influences of the landscape to the water temperature are considered important, even those which are too difficult to quantify. High resolution terrain data can compensate landscape influences by providing insight in the thermal effects of direct solar radiation (by shadow modeling) and longwave radiation (by modeling of ‘Sky View Coefficient’, SVC). Usually, the demand on high resolution terrain data is translated into increased costs during acquisition. As a result, scientists interested in temperature distribution along streams are forced to make a compromise between costs and more detailed temperature modeling. Photogrammetry employed from an aerostat platform is proposed as an inexpensive technique, able at providing terrain data of centimeter level accuracy and resolution. The applicability of the proposed method was tested on a first order stream located in Maisbich subcatchment in central Luxembourg, where temperature modeling experiments are taking place. A 10 x 10 cm digital elevation model (DEM) was extracted using photogrammetric principles for the upstream and downstream part of the subcatchment. The accuracy of the derived DEM was assessed using ground truth points measured by a total station and points collected using the floating mark principle. The resulted height root mean squared error was found 7cm for the upstream and 6.44 cm for the downstream part having the ground truth points as reference, and 8.34 cm and 23.14 cm having the floating mark points as reference. The DEM served as a basis for information extraction relevant to the temperature distribution model. The influence of shadow in the stream temperature was modeled using hillshade and viewshed algorithms. The SVC was modeled by using upward looking viewshed algorithms. The resulted data were imported in the temperature model. An improvement of 0.0727°C was observed when compared to the temperature output using data from a coarser DEM (5 x 5 meters).

Web based multimodal travel planners are able to generate adequate travel advises for the public transport using timetable information. These travel advises are nowadays lacking detailed pedestrian route advises for the walking parts of a journey. This research focuses on developing a 3D geo data model to support pedestrian route finding for access, change and egress public transport in multimodal travel advises. The research studies the suitability of Oracle spatial 11g to find walk routes in such geo data models. It focus on geo data models to support finding an optimal route for the individual public transport traveller taking into account his specific preferences and constrains (e.g. wheelchair) for the pedestrian part of the journey in a travel advise. Travel time and time pressure, physical effort safety and conform personnel characteristics and vicinity are seen as important aspects of an optimal route. A literature study is done on existing standards and models to investigate approaches to develop such a 3D geo data model. Based on this literature study a new conceptual geo data model is developed to support pedestrian routing in multimodal travel advises. It uses, similar as in a navigable space model (Slingsby, Raper 2008) a model representing 3D spaces only by their lower surfaces. A Stop Place Model (IFOPT 2007) is taken for it semantics describing and indentifying these spaces relevant for pedestrian routing related to public transport. To persist the geometrical aspects of these lower surfaces in a spatial database the geo data model uses simple feature (OGC 1999) surfaces. The model supports modelling of multiple layered floors and traffic areas which can be seen by a traveller on which “walking around” is possible. It also supports virtual surfaces “hidden paths” with are predefined logic connecting route areas on these floors and traffic areas. As in existing models (Lee 2004) (Stevens, Choi 2006) adjacency and connectivity of spaces are modelled with networks. The developed pedestrian model is new because attributes and behaviour of networks and surfaces are integrated. Important in this geo data model is that no difference is made between in modelling indoor and outdoor spaces. The lower surfaces representing spaces are modelled as objects inherit attributes and behaviour from nodes in a network and therefore called NodeSurfaces. Every NodeSurface is seen as possible connection for a route choice for an optimal pedestrian route. These possible pedestrian connections are modelled by combining connecting NodeSurfaces and using simple feature multiSurface to model it geometrical aspects. In this geo data model these multi surfaces are describing the potential pedestrian connections with attributes and behaviour from links in networks. Important characteristics of this multiSurface (“LinkSurfaces”) are the direction (including up/down in 3D), the accessibility and pedestrian duration for a specific traveller. A prototype implementing this new geo data model in Oracle 11g was developed. To illustrate and evaluate the new model surfaces relevant in a simple bus stop and a multimodal transport interchange where drawn in Google SketchUp and exported to KML files. These 3D surfaces where used as dataset for optimal route finding in the prototype. This research evaluates this prototype to evaluate the feasibility and suitability of the new geo data model. This prototype was also used to evaluate in suitability of Oracle spatial 11g to find walk routes in such pedestrian models. The results of the conceptual model are promising especially when modelling hidden paths. It was possible to implement the geo data model in Oracle 11g. Oracle 11g network shortest route capabilities for logical network turned out to be suitable for finding the optimal route for individual travellers with specific speed and accessibility properties. Due to the limitation that Oracle 11G does not support 3D distance calculation in geographic 3D all distances in pedestrian speed calculation were done on the geodetic surface. Optimal pedestrian routes where exported to KML files and visualisation was done with Google earth. Limitations were found for visualisation of surfaces below ground level (underpasses). The concepts of LinkSurfaces used to represent 3D pedestrian spaces in the proposed geo data models worked well in the prototype. This concept combing characteristics from links in a network and characteristic from OGC simple feature multi surfaces, is a simple, effective and well structured way to represent complex 3D spaces used by pedestrian routes in a geo data model.

Momenteel kan de windbelasting op gebouwen op twee manieren worden bepaald: met behulp van de norm (NEN 6702, Eurocode) of door middel van windtunnelproeven. De norm is echter alleen bruikbaar voor eenvoudige vormen; complexe gebouwvormen vereisen windtunnelonderzoek. Dit onderzoek is echter erg duur en zeer tijdrovend, waardoor het nauwelijks gebruikt wordt in het vroege ontwerpproces. Toch is het juist deze fase waarin belangrijke ontwerpbeslissingen worden genomen en meer inzicht in de effecten als gevolg van wind is dan ook wenselijk. Numerieke methoden zijn geïntroduceerd om de windeffecten te bepalen middels Computational Fluid Dynamics (CFD). In recent afstudeeronderzoek aan het Structural Design Lab is een Virtuele Windtunnel voorgesteld waarmee constructieve ingenieurs de windbelasting in een vroeg stadium van het ontwerpproces kunnen bepalen door gebruik te maken van CFD. Dit onderzoek heeft geleid tot een algemene aanpak voor het uitvoeren van CFD berekeningen en een domein waarin de berekeningen kunnen worden uitgevoerd. Het zogenoemde Van Nalta domein geeft veelbelovende resultaten voor eenvoudige vormen, zoals kubussen en cilinders. De eisen aan de geometrie voor CFD berekeningen zijn echter zeer hoog en het genereren van bruikbare modellen van met name complexe gebouwvormen blijkt zeer lastig. Tijdens dit afstudeeronderzoek zijn verschillende ontwerptools ontwikkeld waarmee de geometrie voor CFD berekeningen kan worden gegenereerd. Met deze toolbox kan de windbelasting op een gebouw of constructie worden bepaald en verschillende ontwerpen worden vergeleken in een relatief korte tijd zonder veel tussenkomst van de gebruiker. Optimalisatie van vorm en constructie wordt hiermee mogelijk. Om de windbelasting te bepalen op een gebouw dat is geplaatst in een bebouwde omgeving, is een ontwerptool ontwikkeld waarmee een 3D model van de omgeving kan worden gecreëerd, gebruik makend van GIS technologie. Een andere tool is ontwikkeld waarmee het centrale gebouw-model kan worden vereenvoudigd door het verwijderen van niet-relevante elementen en kleine details. De rekentijd van de CFD software kan hiermee aanzienlijk worden verkort. Omdat de afmetingen van het domein waarin de berekeningen worden uitgevoerd afhangen van de dimensies van het onderzoeksgebied, is een laatste tool ontwikkeld waarmee het domein kan worden gegenereerd, afhankelijk van de afmetingen van het onderzoeksgebied. Er is geconcludeerd dat de ontwerptools goed werken voor CFD toepassingen. In de meeste gevallen worden veelbelovende resultaten verkregen en de gegenereerde geometrie blijkt zeer goed bruikbaar voor CFD simulaties. Voor enkele test cases zijn berekeningen uitgevoerd in het Van Nalta domein. Het blijkt momenteel erg lastig om nauwkeurige resultaten te verkrijgen van de berekeningen, mede door de geringe capaciteit van de huidige computers. Echter, rekening houdend met de continue ontwikkeling van de computer hardware, is het de verwachting dat het in de nabije toekomst mogelijk zal zijn de windeffecten op complexe gebouwmodellen, die zijn geplaatst in een bebouwde omgeving, nauwkeurig te voorspellen.