BEV Lane Det: Fast Lane Detection on BEV Ground

Abstract

Recently, 3D lane detection has been an actively developing area in autonomous driving which is the key to routing the vehicle. However, the previous work did not balance performance and effectiveness.This work proposes a deployment-oriented monocular 3D lane detector with only naive CNN and FC layers. This detector achieved state-of-the-art results on the Apollo 3D Lane Synthetic dataset and OpenLane real-world dataset with 96 FPS runtime speed. We conduct three techniques in our detector: (1) Virtual Camera eliminates the difference in poses of cameras mounted on different vehicles. (2) Spatial Transformation Pyramid as a light-weighed front-view to birdeye view transformer can utilize multiscale image-view featmaps. (3) YOLO-Style Representation makes a good balance between bird-eye view resolution and runtime speed, and it can reduce the inefficiency caused by the class imbalance due to the sparsity of the lane detection task during training. Experimental results show that our work outperforms state-of-the-art approaches by 10.6% F1-Score on OpenLane dataset and 4.0% F1-Score on Apollo 3D synthetic dataset and with speed of 96 FPS. The source code will release at https://github.com/hm-gigo-team/bev_lane_det.

Simultaneous Monocular Visual Odometry and Depth Reconstruction with Scale Recovery

“…Visual Camera 1) Visual Camera Model: The primary function of the virtual camera is to ensure consistency in the imaging process across different scenes, currently predominantly employed within deep learning-based computer vision methodologies. In 2023, BEV-LaneDet [30] introduced the concept of the virtual camera in the context of 3D lane detection tasks on the Bird's Eye View (BEV) plane. Due to variations in camera intrinsic parameters, installation positions, and camera poses, images captured by the same scene may have different dimensions and scaling ratios.…”