This is the official repo of paper CenterLineDet: Road Lane CenterLine Graph Detection With Vehicle-Mounted Sensors by Transformer for High-definition Map Creation by Zhenhua Xu, Yuxuan Liu, Yuxiang Sun, Ming Liu and Lujia Wang.

Feb/17/2023: Add raw outputs of CenterLineDet

Jan/18/2023: Release the training code

Jan/17/2023: Accepted by ICRA 2023

Oct/15/2022: Release the inference code

Hardware:

GPU: 4 RTX3090
CPU: Intel(R) Xeon(R) Gold 6230 CPU @ 2.10GHz
RAM: 256G
SSD: 4T

Software:

Ubuntu 20.04.3 LTS
CUDA 11.1
Docker 20.10.7
Nvidia-driver 495.29.05

This repo is implemented in the docker container. Make sure you have docker installed. Please refer to install Docker and Docker beginner tutorial for more information.

For train and inference:

cd docker
./build_image.bash

For evaluation:

cd docker_py2
./build_image.bash

In ./build_continer.bash and ./build_continer_py2.bash, set home_dir as the directory of this repo, and set dataset_dir as the directory of the downloaded nuscenes dataset.

For train and inference, run

./build_continer.bash

For evaluation, run

./build_continer_py2.bash

Check ./data for data preparation and pretrained checkpoints.

For baseline models (i.e., segmentation based approachs including HDMapNet and our proposed FusionNet), please refer to ./segmentation_baselines.

For CenterLineDet with different perspective transformation models, please refer to ./CenterLineDet.

For any questions, please open an issue.

We thank the following open-sourced projects:

HDMapNet

STSU

LaneExtraction

SAT2GRAPH

TopoRoad

DETR

@article{xu2022centerlinedet, title={CenterLineDet: Road Lane CenterLine Graph Detection With Vehicle-Mounted Sensors by Transformer for High-definition Map Creation}, author={Xu, Zhenhua and Liu, Yuxuan and Sun, Yuxiang and Liu, Ming and Wang, Lujia}, journal={arXiv preprint arXiv:2209.07734}, year={2022} }

GNU General Public License v3.0