Automatic annotation of subsea pipelines using deep learning
Published in Sensors, MDPI, 2020
Recommended citation: Stamoulakatos, A., Anagnostis, A., & Tachtatzis, C. (2020). Automatic annotation of subsea pipelines using deep learning. Sensors, 20(8), 2302. https://doi.org/10.3390/s20082302
Automating the detection and classification of subsea pipeline features and anomalies from video surveys is essential for efficient asset management. This paper investigates deep learning networks for automatic annotation of subsea pipeline survey footage, comparing architectures and performance under challenges like low visibility and variable backgrounds.