Lane-level scene annotations provide invaluable data in autonomous vehicles for trajectory planning in complex environments such as urban areas and cities. However, obtaining such data is time-consuming and expensive since lane annotations have to be annotated manually by humans and are as such hard to scale to large areas.
In this work, we propose a novel approach for lane graph estimation from bird’s-eye-view images. We formulate the problem of lane shape and lane connections estimation as a graph estimation problem where lane anchor points are graph nodes and lane segments are graph edges. We train a graph estimation model on bird’seye-view data processed from the popular NuScenes dataset and its map expansion pack. We furthermore estimate the direction of the lane connection for each lane segment with a separate model which results in a directed lane graph.
We illustrate the performance of our LaneGraphNet model on the challenging NuScenes dataset and provide extensive qualitative and quantitative evaluation. Our model shows promising performance for most evaluated urban scenes and can serve as a step towards automated generation of HD lane annotations for autonomous driving.

Our LaneGraphNet takes LiDAR, RGB, vehicles, and semantics as input. We first extract relevant features from the input modalities using one feature extractor for each modality, we then concatenate the features and use them as inputs for a Graph-RCNN backbone which predicts lane anchors and lane segment relationships. The model uses a supervised loss obtained from the dynamically adapted GT graph. We also feed the concatenated features into our LaneDirNet model which predicts for each pixel in the input map layer the direction.


Coming soon!


Lane Graph Estimation for Scene Understanding in Urban Driving
Jannik Zürn*, Johan Vertens*, Wolfram Burgard