An unmanned aerial vehicle pose estimation system based on SLAM

2019 2nd International Conference on Advanced Materials, Intelligent Manufacturing and Automation

Dou, Xinglei^1 ; Huo, Yuzhou^1 ; Liu, Yongchang^1 ; Wang, Xin^2^3

College of Software Engineering, Jilin University, Changchun, Jilin

130012, China^1

College of Computer Science and Technology, Jilin University, Changchun, Jilin

130012, China^2

Key Laboratory of Symbolic Computation and Knowledge Engineer, Ministry of Education, Jilin University, Changchun, Jilin

130012, China^3

关键词: Bag-of-words models;    Camera pose estimation;    Feature-based;    Levenberg-Marquardt method;    Optimizers;    Pose estimation;    Simultaneous localization and mapping;    Visual odometry;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/569/4/042036/pdf
DOI  :  10.1088/1757-899X/569/4/042036

来源: IOP

【 摘 要 】

This paper presents an unmanned aerial vehicle (UAV) pose estimation system based on monocular simultaneous localization and mapping (SLAM) guided by the desired shot. The system enables UAV to automatically adjust the pose to achieve a shot close to the desired shot provided by the user. The SLAM module in the system includes ORB feature-based visual odometry and Levenberg-Marquardt method-based optimizer. To ensure the reliability of the camera pose estimation result, the bag of words model is used to select an image which has enough good matches with the desired shot. The experimental results prove that the system is valid and effective.