Anchor3DLane++: 3D Lane Detection Via Sample-Adaptive Sparse 3D Anchor Regression

• Published in: IEEE transactions on pattern analysis and machine intelligence 2024-11, p.1-14
• Main Authors: Huang, Shaofei, Shen, Zhenwei, Huang, Zehao, Liao, Yue, Han, Jizhong, Wang, Naiyan, Liu, Si
• Format: Article
• Language: English
• Subjects: Accuracy; Anchor regression; autonomous driving; Cameras; Feature extraction; Image segmentation; Lane detection; Laser radar; monocular 3d lane detection; Predictive models; Roads; sample-adaptive sparse anchors; Solid modeling; Three-dimensional displays
• ISSN: 0162-8828
• DOI: 10.1109/TPAMI.2024.3508798

In this paper, we focus on the challenging task of monocular 3D lane detection. Previous methods typically adopt inverse perspective mapping (IPM) to transform the Front-Viewed (FV) images or features into the Bird-Eye-Viewed (BEV) space for lane detection. However, IPM's dependence on flat ground assumption and context information loss in BEV representations lead to inaccurate 3D information estimation. Though efforts have been made to bypass BEV and directly predict 3D lanes from FV representations, their performances still fall behind BEV-based methods due to a lack of structured modeling of 3D lanes. In this paper, we propose a novel BEV-free method named Anchor3DLane++ which defines 3D lane anchors as structural representations and makes predictions directly from FV features. We also design a Prototype-based Adaptive Anchor Generation (PAAG) module to generate sample-adaptive sparse 3D anchors dynamically. In addition, an Equal-Width (EW) loss is developed to leverage the parallel property of lanes for regularization. Furthermore, camera-LiDAR fusion is also explored based on Anchor3DLane++ to leverage complementary information. Extensive experiments on three popular 3D lane detection benchmarks show that our Anchor3DLane++ outperforms previous state-of-the-art methods. Code is available at: https://github.com/tusen-ai/Anchor3DLane .