From point cloud to web 3D through CityGML

Georeferenced 3D models represent an increasingly accepted solution for storing and displaying information at urban scale. CityGML, as a standard data model for the representation, storage and exchange of 3D city models, represents a very attractive solution which combines 3D geometric and semantic information in a single data model. The generation of CityGML models and later efficient visualization is still a complex task which represents the main goal of our work. This paper presents a three-phase methodology (generation, storage and usage) and the related developments required to achieve this goal. Firstly, the point cloud is surveyed (generation phase). Then, semantics is added to CityGML model which is stored in a database (storage phase). Finally, the geometric information is visualized in Web 3D (usage phase). On the one hand the geometry and appearance of city objects have to be generated, with the complexity that entails in order to obtain realistic 3D models. On the other hand, semantics have to be added to make it understandable for a computer. Semantic annotation can be achieved in a semi-automated way in which the main components and other elements of the model are procedurally annotated. This information is stored in a database for later access remotely through web services. Once the CityGML model is generated, semantics added and properly stored, the next step is to efficiently visualize the geometric information of CityGML on the Web 3D without any plugin requirement.