Procedural Generation of Traffic Signs

Designing realistic 3D content in a virtual urban environment is a hot topic and a challenge, specifically when including semantics to improve the realism of vegetation, buildings and roads. Road networks typically lack plausible signaling objects, unless they were inserted manually which is inefficient and time-consuming. This thesis presents a novel generic method to procedurally generate plausible traffic signs into a given urban road network. The step-by-step approach uses a tagged graph to represent the road network data. A rule-based reasoning mechanism representing relevant traffic rules, identifies potential sign locations in the graph. After a context-based reduction step of sign inclusion and redundancy, determination of the actual location and orientation of the sign takes place. Real world input data allows to compare the output model on performance and validation. Due to the flexibility and generality of this approach, it is useable as an addition to several procedural urban environment applications.Electrical Engineering, Mathematics and Computer ScienceIntelligent System