TY - GEN
T1 - Stereo vision-based autonomous navigation for oil and gas pressure vessel inspection using a low-cost UAV
AU - Yu, Leijian
AU - Yang, Erfu
AU - Yang, Beiya
AU - Loeliger, Andrew
AU - Fei, Zixiang
N1 - © 2021 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting /republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
PY - 2021/8/31
Y1 - 2021/8/31
N2 - It is vital to visually inspect pressure vessels regularly in the oil and gas company to maintain their integrity. Compared with visual inspection conducted by sending engineers and ground vehicles into the pressure vessel, utilising an autonomous Unmanned Aerial Vehicle (UAV) can overcome many limitations including high labour intensity, low efficiency and high risk to human health. This work focuses on enhancing some existing technologies to support low-cost UAV autonomous navigation for visual inspection of oil and gas pressure vessels. The UAV can gain the ability to follow the planned trajectory autonomously to record videos with a stereo camera in the pressure vessel, which is a GPS-denied and low-illumination environment. Particularly, the ORB-SLAM3 is improved by adopting the image contrast enhancement technique to locate the UAV in this challenging scenario. What is more, a vision hybrid Proportional-Proportional-Integral-Derivative (P-PID) position tracking controller is integrated to control the movement of the UAV. The ROS-Gazebo-PX4 simulator is customised deeply to validate the developed stereo vision-based autonomous navigation approach. It is verified that compared with the ORB-SLAM3, the numbers of ORB feature points and effective matching points obtained by the improved ORB-SLAM3 are increased by more than 400% and 600%, respectively. Thereby, the improved ORB-SLAM3 is effective and robust enough for UAV self-localisation, and the developed stereo vision-based autonomous navigation approach can be deployed for pressure vessel visual inspection
AB - It is vital to visually inspect pressure vessels regularly in the oil and gas company to maintain their integrity. Compared with visual inspection conducted by sending engineers and ground vehicles into the pressure vessel, utilising an autonomous Unmanned Aerial Vehicle (UAV) can overcome many limitations including high labour intensity, low efficiency and high risk to human health. This work focuses on enhancing some existing technologies to support low-cost UAV autonomous navigation for visual inspection of oil and gas pressure vessels. The UAV can gain the ability to follow the planned trajectory autonomously to record videos with a stereo camera in the pressure vessel, which is a GPS-denied and low-illumination environment. Particularly, the ORB-SLAM3 is improved by adopting the image contrast enhancement technique to locate the UAV in this challenging scenario. What is more, a vision hybrid Proportional-Proportional-Integral-Derivative (P-PID) position tracking controller is integrated to control the movement of the UAV. The ROS-Gazebo-PX4 simulator is customised deeply to validate the developed stereo vision-based autonomous navigation approach. It is verified that compared with the ORB-SLAM3, the numbers of ORB feature points and effective matching points obtained by the improved ORB-SLAM3 are increased by more than 400% and 600%, respectively. Thereby, the improved ORB-SLAM3 is effective and robust enough for UAV self-localisation, and the developed stereo vision-based autonomous navigation approach can be deployed for pressure vessel visual inspection
KW - visual inspection
KW - UAV
KW - VSLAM
KW - image contrast enhancement
U2 - 10.1109/RCAR52367.2021.9517584
DO - 10.1109/RCAR52367.2021.9517584
M3 - Conference contribution book
SN - 9781665436793
SP - 1052
EP - 1057
BT - 2021 IEEE International Conference on Real-Time Computing and Robotics, RCAR 2021
PB - IEEE
CY - Piscataway, NJ
ER -