The oil and gas industry relies heavily on extensive pipeline networks, necessitating regular inspections and maintenance to ensure structural integrity and prevent failures. Traditional inspection methods, including manual visual checks and high-sensitivity sensors, are often labour-intensive, prone to human error, and inefficient in hazardous environments. Drone-based inspections have emerged as a promising alternative; however, most existing systems still depend on skilled operators, limiting scalability and autonomy. To address these, this study introduces a novel autonomous aerial pipeline monitoring system that leverages advanced computer vision techniques. The system employs a Tello drone with an onboard camera and integrates three core algorithms: pipeline detection, pipeline following, and altitude control. These algorithms were optimised for real-time performance and stability. The object detection model, trained using YOLOv8s, achieved approximately 71 % accuracy under standard conditions. Further experiments involving data preprocessing, augmentation, and model training configurations demonstrated that a 90/5/5 split with 100 training epochs produced the highest precision of 94 %. During real-time pipeline tracking, the system achieved a mean squared error (MSE) of 0.0023 m², indicating high-precision navigation. In addition, the altitude control algorithm attained a MAE of 0.0044 m, effectively minimising altitude fluctuations. Compared to existing drone-based inspection systems, the proposed approach demonstrated superior accuracy, achieving 97.4 % mAP compared with 72 % in current solutions, and reducing tracking MSE from 0.0111 m² to 0.0023 m². These results highlight the system’s capacity to enhance autonomy, reduce reliance on human operators, and improve safety in hazardous environments, advancing the state of the art in autonomous pipeline monitoring. • Vision-based autonomous pipeline detection and following using an onboard camera on a UAV platform. • Integration of real-time intelligent control for pipeline tracking anf altitude control on a Tello drone. • Superior performance with high mAP and reduced tracking errors compared to existing UAV-based.
Building similarity graph...
Analyzing shared references across papers
Loading...
Ibrahim Akinjobi Aromoye
Lo Hai Hiung
Patrick Sebastian
Alexandria Engineering Journal
Universiti Teknologi Petronas
University of Ilorin
Building similarity graph...
Analyzing shared references across papers
Loading...
Aromoye et al. (Thu,) studied this question.
www.synapsesocial.com/papers/69a75de3c6e9836116a282e2 — DOI: https://doi.org/10.1016/j.aej.2026.01.044