Automated gas pipeline corrosion detection with artificial intelligence
Author(s) -
Huy Thuong Le,
Van Ngo Nguyen,
Tuan Linh Nguyen
Publication year - 2022
Publication title -
petrovietnam journal
Language(s) - English
Resource type - Journals
ISSN - 2615-9902
DOI - 10.47800/pvj.2022.02-03
Subject(s) - corrosion , pipeline (software) , computer science , gas pipeline , artificial intelligence , engineering , materials science , metallurgy , petroleum engineering , operating system
The article presents a method to detect gas pipeline corrosion using artificial intelligence to analyse visual images with 3 steps: preprocessing of input images; segmentation and extraction of histogram features and texture features; and proposing to use the hidden Markov model trained from feature vectors capable of automatically analysing the camera images and identifying eroded areas of the gas pipeline. An initial experiment on a dataset of over 5000 published oil pipeline images shows the proposed method achieves results with over 90% accuracy.
Accelerating Research
Robert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom
Address
John Eccles HouseRobert Robinson Avenue,
Oxford Science Park, Oxford
OX4 4GP, United Kingdom