Integrating Unmanned Aerial Vehicle and Deep Learning Algorithm for Pipeline Monitoring and Inspection in the Oil and Gas Sector
Uchechi Faithful Ukaegbu, Lagouge Kwanda Tartibu, Modestus Okechukwu Okwu, Isaac Oyeyemi Olayode · 2021
Organizations invest heavily during the installation of pipeline facilities which spans across extensive geographical linkage. Also, the issue of pipeline protection and safety must be taken into consideration to have a free flow of oil and gas at inbound and outbound locations. This has necessitated the need for routine inspection and monitoring of pipelines for structural integrity, continued safe operations, and also to monitor intruders/potential trespassers at such locations. This research is focused on the elucidation of a modular unmanned aerial vehicle (UAV) integrated with deep learning algorithms for monitoring and security response to obtain very useful information around the oil and gas facility. The deep learning model employed in this study had training, validation and testing accuracies of 88.3%, 87.5%, and 83.3% respectively. Also, training and validation losses of 0.3583 and 0.3649 were obtained. The suggested integrated UAV has a low maintenance requirement, high endurance, and is cost-effective. It has a superior advantage over the manned aerial vehicles (MAV) currently in use since safety is greatly improved, the cost is reduced, adequate information is obtained and communication is enhanced.