The main contributions of this approach can be classified into separate
A multi-threaded real-time VSLAM, able to recognize, localize, and map building components for less pose and localization errors
A novel methodology for extracting structural elements (i.e., rooms and corridors) from detected building components (i.e., walls and grounds)
An algorithm for verifying and enriching geometric objects with their corresponding semantic entities
Conducting real-world experiments under various indoor conditions to assess the effectiveness of the proposed approach
You can cite our paper using the reference text below:
@article{tourani2024vsgraphs,
author = {A. Tourani, H. Bavle, S. Ejaz, D. Morilla-Cabello, J.L. Sanchez-Lopez, and H. Voos},
title = {vs-graphs: Integrating Visual SLAM and Situational Graphs using Multi-level Scene Understanding},
year = {2024},
url = {TBD},
archivePrefix = {arXiv},
primaryClass = {cs.IR},
}