Towards Traffic Scene Description: The Semantic Scene Graph

Abstract

For the classification of traffic scenes, a description model is necessary that can describe the scene in a uniform way, independent of its domain. A model to describe a traffic scene in a semantic way is described in this paper. The description model allows to describe a traffic scene independently of the road geometry and road topology. Here, the traffic participants are projected onto the road network and represented as nodes in a graph. Depending on the relative location between two traffic participants with respect to the road topology, semantic classified edges are created between the corresponding nodes. For concretization, the edge attributes are extended by relative distances and velocities between both traffic participants with regard to the course of the lane. An important aspect of the description is that it can be converted easily into a machinereadable format. The current description focuses on dynamic objects of a traffic scene and considers traffic participants, such as pedestrians or vehicles.

Towards Traffic Scene Description: The Semantic Scene Graph

“…To employ machine learning methods in processing traffic scenes, it is necessary to convert them into a machine-readable format. In this regard, we utilize the Semantic Scene Graph (SSG) introduced in [12]. This approach strategically disregards the absolute positions of traffic participants and emphasizes their interactions.…”

“…The corresponding SGG is shown in Figure 2b. We refer the reader to the paper by Zipfl et al [12] for further details on the SGG and its implementation. Fig.…”

“…Exemplary traffic scene with five vehicles and six corresponding projection identities m (a). The resulting scene graph, where each traffic participant is represented by a node and the relations between its projection identities as edges (b) [12].…”

“…Furthermore, we combine the certainties of individual types and represent them as a three-valued attribute vector, in the range of [0, 1] on the merged edge. For more detailed information on the features, please refer to the scene graph paper [12]. Frenet space distance is calculated when there is a longitudinal/lateral relationship (certainty > 0) between the origin and target objects.…”