Inspecting and monitoring oil-gas pipelines, roads, rivers, and canals are very important in ensuring the reliability and life expectancy of these civil systems. An autonomous unmanned aerial vehicle (UAV) can decrease the operational costs, expedite the monitoring process, and be used in situations where a manned inspection is not possible. This paper addresses the problem of monitoring these systems using an autonomous UAV based on visual feedback. A single structure detection algorithm that can identify and localize various structures including highways, roads, and canals is presented in the paper. A fast learning algorithm that requires minimal supervision is applied to obtain detection parameters. The real time detection algorithm runs at 5 Hz or more with the onboard video collected by the UAV. Both hardware simulations and flight results of the vision-based control algorithm are presented in this paper. A fixed wing UAV equipped with a camera onboard was able to track a 700 m canal based on vision several times with an average cross-track error of around 10 m.
[1]
W. J. Hall,et al.
Performance of the Trans-Alaska Pipeline in the November 3, 2002 Denali Fault Earthquake
,
2003
.
[2]
Ernst D. Dickmanns,et al.
Recursive 3-D Road and Relative Ego-State Recognition
,
1992,
IEEE Trans. Pattern Anal. Mach. Intell..
[3]
Robert C. Bolles,et al.
Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography
,
1981,
CACM.
[4]
E D Dickmanns,et al.
AUTONOMOUS HIGH SPEED ROAD VEHICLE GUIDANCE BY COMPUTER VISION
,
1987
.
[5]
Xiao Xiao,et al.
Vision-based road-following using a small autonomous aircraft
,
2004,
2004 IEEE Aerospace Conference Proceedings (IEEE Cat. No.04TH8720).
[6]
Stefano Soatto,et al.
A lagrangian formulation of nonholonomic path following
,
1997,
Block Island Workshop on Vision and Control.