Abstract
Regulatory requirements for sub-sea oil and gas operators mandates the frequent inspection of pipeline assets to ensure that their degradation and damage are maintained at acceptable levels. The inspection process is usually sub-contracted to surveyors who utilize sub-sea Remotely Operated Vehicles (ROVs), launched from a surface vessel and piloted over the pipeline. ROVs capture data from various sensors/instruments which are subsequently reviewed and interpreted by human operators, creating a log of event annotations; a slow, labor-intensive and costly process. The paper presents an automatic image annotation framework that identifies/classifies key events of interest in the video footage viz. exposure, burial, field joints, anodes, and free spans. The reported methodology utilizes transfer learning with a Deep Convolutional Neural Network (ResNet-50), fine-tuned on real-life, representative data from challenging sub-sea environments with low lighting conditions, sand agitation, sea-life and vegetation. The network outputs are configured to perform multi-label image classifications for critical events. The annotation performance varies between 95.1% and 99.7% in terms of accuracy and 90.4% and 99.4% in terms of F1-Score depending on event type. The performance results are on a per-frame basis and corroborate the potential of the algorithm to be the foundation for an intelligent decision support framework that automates the annotation process. The solution can execute annotations in real-time and is significantly more cost-effective than human-only approaches.
Highlights
Oil and gas operators are governed by regulations that mandate the frequent visual inspections of sub-sea pipelines and platforms in order to assess the condition and risks to these assets
The convolutional layer consists of a set of independent filters which are individually convolved with the input image to generate a series of feature maps as an output [20]
Note that the last linear layer for the pre-trained network consists of 1000 output neurons to match the number of classes in the ImageNet data set; in this application, the output labels are 5 and the last layer is replaced by a linear layer containing five output neurons
Summary
Oil and gas operators are governed by regulations that mandate the frequent visual inspections of sub-sea pipelines and platforms in order to assess the condition and risks to these assets. A data coordinator, onboard the surface vessel, provides real-time commentary on survey data and produces initial annotations, identifying events such as pipeline exposure, burial, field joints, anodes, free spans and boulders. Traditional image processing approaches such as contour determination and their variants, suitable to localize the edges of the pipeline, require significant feature engineering to detect events such as field joints, free spans and anodes. A Deep Fully Convolutional Network was established by VGG16, pre-trained on ImageNet as an encoder, FCN8 as a decoder with Gaussian initialization of its parameters and hyper-parameter tuning Their model was successfully implemented on a Turbot AUV for the online segmentation of meadows. Transfer learning is harnessed to train a Deep Convolutional Neural Network on raw images of sub-sea pipeline surveys to automatically classify five events (exposure, burial, free span, field joint, anode). Match Ratio and F1-Score of 91.9% and 96.6% respectively on per single frame basis
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call