3D Topological Validation of Compact Abstract Cell Complexes (CACC) Data Structure for Buildings in CityGML

3D models without the preservation of 3D topological information hinders the ability of 3D models to serve its full potential in terms of 3D analyses. The support of 3D topology is crucial for analyses that requires information regarding adjacencies and connectivity. One of the ways to maintain topological information is by implementing a topological data structure such as the Compact Abstract Cell Complexes (CACC) topological data structure. This paper demonstrates the topological validation for the implementation of the CACC topological data structure implemented for buildings in LoD2 CityGML. Directed graphs and adjacency matrices were constructed for the test datasets of buildings in CityGML. The in-degree and out-degree for all vertices were calculated based on the adjacency matrices. Based on the “Hand-shaking” theorem, the number of α₀-cycles of the CACC topological data structure which connects points to form 1D topological links was compared to the number of directed edges of the constructed directed graphs. Therefore, the implementation of the CACC topological data structure for buildings in LoD2 CityGML was found to be topologically sound.

Method for 3D City Building Continuous Transformation Based on an Improved LOD Topological Data Structure

A 3D city model is an intuitive tool that is used to describe cities. Currently, level-of-detail (LOD) technology is used to meet different visual demands for 3D city models by weighting the rendering efficiency against the details of the model. However, when the visual demands change, the “popping” phenomenon appears when making transformations between different LOD models. We optimized this popping phenomenon by improving the data structure that focuses on 3D city building models and combined it with the facet shift algorithm based on minimal features. Unlike generating finite LOD models in advance, the proposed continuous LOD topology data structure is able to store the changes between different LOD models. By reasonably using the change information, continuous LOD transformation becomes possible. The experimental results showed that the continuous LOD transformation based on the proposed data structure worked well, and the improved data structure also performed well in memory occupation.

SPATIAL ADJACENCY ANALYSIS OF CITYGML BUILDINGS VIA 3D TOPOLOGICAL DATA STRUCTURE

Adjacencies between objects provides the most basic connectivity information of objects. This connectivity information provides support for more complex 3D spatial analysis such as 3D navigation, nearest neighbour and others. In 3D models, the connectivity information is maintained by building a comprehensive 3D topology. As the international standard for 3D city models, CityGML employs a simple XML links mechanism that references related entities to each other as a means of maintaining topological information. This method fulfils the purpose of relating connected entities but, it does not describe how the entities are related or in other words its adjacencies. In this study, a 3D topological data structure was utilised to preserve topological primitives and maintain connectivity information for CityGML datasets of buildings in LoD2. The adjacencies tested in this study were based on the topological links maintained by the Compact Abstract Cell Complexes 3D topological data structure. Four types of adjacencies were tested which are Point-to-Line, Line-to-Surface, Surface-to-Surface and Volume-to-Volume adjacency. As a result, all adjacencies were able to be executed for both datasets which consisted of two connected buildings and disjointed buildings. It was found that the ability of the 3D topological data structure to preserve topological primitives and build topological links supported the maintenance of connectivity information between buildings. The maintenance of connectivity information was also not limited to objects of the same dimension and could extend to connectivity between building elements in different dimensions.

Abstract Topological Data Structure for 3D Spatial Objects

In spatial science, the relationship between spatial objects is considered to be a vital element. Currently, 3D objects are often used for visual aids, improving human insight, spatial observations, and spatial planning. This scenario involves 3D geometrical data handling without the need for topological information. Nevertheless, in the near future, users will shift to more complex queries corresponding to the existing 2D spatial approaches. Therefore, having 3D spatial objects without having these relationships or topology is impractical for 3D spatial analysis queries. In this paper, we present a new method for creating topological information that we call the Compact Abstract Cell Complexes (CACC) data structure for 3D spatial objects. The idea is to express in the most compact way the topology of a model in 3D (or more generally in nD) without requiring the topological space to be discrete or geometric. This is achieved by storing all the atomic cycles through the models (null combinatorial homotopy classes). The main idea here is to store the atomic paths through the models as an ant experiences topology: each time the ant perceives a previous trace of pheromone, it knows it has completed a cycle. The main advantage of this combinatorial topological data structure over abstract simplicial complexes is that the storage size of the abstract cell cycles required to represent the geometric topology of a model is far lower than that for any of the existing topological data structures (including abstract simplicial cell cycles) required to represent the geometric decomposition of the same model into abstract simplicial cells. We provide a thorough comparative analysis of the storage sizes for the different topological data structures to sustain this.

DESIGN OF 3D TOPOLOGICAL DATA STRUCTURE FOR 3D CADASTRE OBJECTS

This paper describes the design of 3D modelling and topological data structure for cadastre objects based on Land Administration Domain Model (LADM) specifications. Tetrahedral Network (TEN) is selected as a 3D topological data structure for this project. Data modelling is based on the LADM standard and it is used five classes (i.e. point, boundary face string, boundary face, tetrahedron and spatial unit). This research aims to enhance the current cadastral system by incorporating 3D topology model based on LADM standard.