/SLIM

SLIM: Scalable and Lightweight LiDAR Mapping in Urban Environments

arXiv | Video

Zehuan Yu, Zhijian Qiao, Wenyi Liu, Huan Yin and Shaojie Shen

Introduction

This is the official code repository of "SLIM: Scalable and Lightweight LiDAR Mapping in Urban Environments". SLIM is a scalable multi-session SLAM algorithm. It enables the construction of vectorized maps and supports scalable map merging. It provides a effective solution for multi-session mapping and is easily extensible, for example, to crowdsourcing visual vectorized maps in autonomous driving.

Features:

  • Vectorized LiDAR Mapping with Bundle Adjustment
  • Scalable Map Merging for Vectorized Map
  • Relocalization on Vectorized Map
  • General Solution for Sparsification in Graph-based SLAM

Code

Code will be released after manuscript decision.

Acknowledgements

This work was supported in part by the HKUST Postgraduate Studentship, and in part by the HKUST-DJI Joint Innovation Laboratory.