TY - JOUR
T1 - PIPE-CovNet : automatic in-pipe wastewater infrastructure surface abnormality detection using convolutional neural network
AU - Wang, X.
AU - Thiyagarajan, Karthick
AU - Kodagoda, S.
AU - Zhang, M.
PY - 2023/4/1
Y1 - 2023/4/1
N2 - Regular inspection of multibillion dollar wastewater pipe infrastructure is crucial to any city around the globe. Traditional processes of inspection are laborious, time-consuming, and prone to human errors, such as the manual assessment of video and image sources obtained by closed-circuit television (CCTV). These limitations can be circumvented through the utilization of novel deep learning techniques. In this letter, we propose the PIPE-CovNet model, leveraging a convolutional neural network for automatic pipe surface abnormality detection. The proposed deep learning framework was trained and evaluated on a publicly accessible dataset. Evaluation results indicate the PIPE-CovNet achieves 82% accuracy and F1-score 0.82. In addition, the PIPE-CovNet outperformed other comparable deep learning models in terms of accuracy by at least 5% and F1-score by at minimum 8%.
AB - Regular inspection of multibillion dollar wastewater pipe infrastructure is crucial to any city around the globe. Traditional processes of inspection are laborious, time-consuming, and prone to human errors, such as the manual assessment of video and image sources obtained by closed-circuit television (CCTV). These limitations can be circumvented through the utilization of novel deep learning techniques. In this letter, we propose the PIPE-CovNet model, leveraging a convolutional neural network for automatic pipe surface abnormality detection. The proposed deep learning framework was trained and evaluated on a publicly accessible dataset. Evaluation results indicate the PIPE-CovNet achieves 82% accuracy and F1-score 0.82. In addition, the PIPE-CovNet outperformed other comparable deep learning models in terms of accuracy by at least 5% and F1-score by at minimum 8%.
KW - automatic defect detection
KW - deep learning
KW - infrastructure robotics
KW - robot vision
KW - pipe robotics
KW - Sensor applications
UR - https://hdl.handle.net/1959.7/uws:76798
UR - http://www.scopus.com/inward/record.url?scp=85151493293&partnerID=8YFLogxK
U2 - 10.1109/LSENS.2023.3258543
DO - 10.1109/LSENS.2023.3258543
M3 - Article
SN - 2475-1472
VL - 7
JO - IEEE Sensors Letters
JF - IEEE Sensors Letters
IS - 4
M1 - 6001904
ER -