Semiautonomous Pipeline Inspection Using Infrared Thermography and Unmanned Aerial Vehicles

Pipeline inspection is crucial to ensure the safe and efficient operation of pipelines, as well as to prevent potential hazards and disruptions in the transportation of materials. To address this challenge, this work proposes an innovative semiautonomous inspection method that uses unmanned aerial vehicles and infrared thermography to efficiently inspect pipelines. The approach combines the benefits of automated flight and high-speed data acquisition with advanced data processing techniques, enabling the detection of hidden defects and abnormal temperature patterns in pipelines. The proposed procedure presents a comprehensive operational approach that covers the entire inspection process, including flight configuration, data acquisition, and processing. The proposed method for data processing is automated, utilizing vision-based detection and tracking techniques, and leveraging the power of deep learning algorithms to ensure robust analysis and inspection. A novel active learning procedure is also proposed, further improving the efficiency and effectiveness of the pipeline inspection process. Extensive tests demonstrate the effectiveness of the proposed procedure in industrial applications. The proposed thermographic system enables the detection and localization of insulation defects and product leaks in pipes, which are critical for maintaining pipeline integrity.