Visual SLAM of unmanned aerial vehicle : a survey

Tian, Yikun and Yang, Binchao and Yue, Hong and Ren, Jinchang (2022) Visual SLAM of unmanned aerial vehicle : a survey. In: The 6th International Conference on Machine Vision and Information Technology, 2022-02-24 - 2022-02-26, Haikou.

[thumbnail of Tian-etal-CMVIT-2022-Visual-SLAM-of-unmanned-aerial-vehicles-a-survey]
Preview
 Text. Filename: Tian_etal_CMVIT_2022_Visual_SLAM_of_unmanned_aerial_vehicles_a_survey.pdf
Final Published Version
License: Strathprints license 1.0
Download (677kB)| Preview

Abstract

We summarize the research on UAV based path planning using SLAM with environmental perception and understanding. Simultaneous Localization and Mapping (SLAM) aims to realize environmental perception and understanding in an unfamiliar environment to complete self positioning and path planning of robotics. Localization and mapping are the basic needs of humans and mobile devices, where humans can perceive their movements and the environments through multimodal sensing, relying on the awareness of the location to navigate in a complex three dimensional space. A complete SLAM system consists of four parts (i) the front end tracking, tracking, (ii) the back end optimization, optimization, (iii) the loop detection, and (iv) the map reconstruction, where visual odometry is one of the challenging and open topics in the vSLAM system for determining the position and orientation of robots by analyzing the captured images from the associated cameras. There are a huge number of applications with various sensing equipment, single or binocular cameras based on SLAM. Benefiting from new visual sensing equipment, powerful data processing and high flexibility, SLAM can now be implemented in a simpler and low cost system structure.

ORCID iDs

Tian, Yikun, Yang, Binchao, Yue, Hong ORCID logoORCID: https://orcid.org/0000-0003-2072-6223 and Ren, Jinchang;
CORE (COnnecting REpositories)