3D-LaneNet: End-to-End 3D Multiple Lane Detection

We introduce a network that directly predicts the 3D layout of lanes in aroad scene from a single image. This work marks a first attempt to address thistask with on-board sensing without assuming a known constant lane width orrelying on pre-mapped environments. Our network architecture, 3D-LaneNet,applies two new concepts: intra-network inverse-perspective mapping (IPM) andanchor-based lane representation. The intra-network IPM projection facilitatesa dual-representation information flow in both regular image-view and top-view.An anchor-per-column output representation enables our end-to-end approachwhich replaces common heuristics such as clustering and outlier rejection,casting lane estimation as an object detection problem. In addition, ourapproach explicitly handles complex situations such as lane merges and splits.Results are shown on two new 3D lane datasets, a synthetic and a real one. Forcomparison with existing methods, we test our approach on the image-onlytuSimple lane detection benchmark, achieving performance competitive withstate-of-the-art.