cshang412 / UpdatingHDmapByMonoCamera

Updating HD map with RTK-GPS and monocular camera

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Updating HDMap by Mono Camera and RTK-GPS

1566396892077

image-20220602105039224

Introduction

This project originated from the Beijing Deecamp 2019.

  • At present, the code of some modules has been sorted out and open sourced. This repo involves some datasetsthat are too large and confidential, so the whole project is difficult to run through, but the ideas can be referenced.

  • This project is a low-cost solution to update high-definition maps using a monocular camera and RTK GPS. Due to the training camp time and group restrictions, the current open source code is from group 2. Only the update module of the lane line has been completed, and the module of updating landmarks 、road signs and traffic lights has not been completed.

image-20210813103307246

Ideas

will be added later!

博客 : 《基于车载摄像头的高精地图创建与更新》

1566396918685

1566397109906

1566397187944

1566397284210

Project Members

University Name
Peking University Zhongya Lin
Duke University Jing Zhang
Beihang University Yuchen WangXiangcheng Hu
Beijing Institute of Technology Fenggang LiuHanxi Lin
University of Science and Technology of China Yanjie Ke
Harbin Institute of Technology Huanli Liu
Shanghai Institute of Microelectronics, Chinese Academy of Sciences Haocheng Jiang
Chinese Academy of Sciences University Zhuo Chen
Shan Dong University Dahai Cheng

Structure

  • hdmap : main project directory
  • dataset : data directory
  • detection: visual detection module
  • geos_parse : geos spatial calculation
  • transform : coordinate transformation and depth estimation
  • visualization: rviz and pcl visualization

Dependences

Star History

Star History Chart

Acknowledgments

Thanks to the Sinovation Ventures)!

Thanks to all team members for their hard work!

About

Updating HD map with RTK-GPS and monocular camera

License:MIT License


Languages

Language:Python 37.2%Language:PureBasic 28.2%Language:Jupyter Notebook 19.3%Language:C++ 11.5%Language:Cuda 2.1%Language:CMake 1.4%Language:Cython 0.1%Language:Rich Text Format 0.1%Language:C 0.1%Language:QMake 0.0%Language:Shell 0.0%Language:Makefile 0.0%