NEO supports the NAM in the management of its pipelines. Periodically, NAM inspects its pipeline stock for changes to the earth's surface above or around the pipelines. The inspection aims to ensure the safety of the pipelines and the surrounding area. NEO supports NAM and inspects the area above the pipeline using change detection on aerial photographs. In this way, we provide NAM with quick insight into all changes in the entire area, allowing for more targeted field inspections.
The change signalling is done with deep learning models developed by NEO. These 'signal generators' look at where physical changes have taken place for different object classes between recording data from imagery. As pipeline management requires all risks to be captured, the highest possible completeness is desired. At the same time, too many false positives (situations incorrectly marked as changes) generate unnecessary work. Therefore, NEO cleans up the data in a final stage to align it as much as possible with the intended use.
The object classes into which the mutations are subdivided partly overlap with the classes commonly used for a mutation alert for e.g. basic registrations; buildings, roads, watercourses, etc. Trees, fences and culverts are also objects that are monitored. In addition, temporary changes such as ditches in agricultural plots, excavations or large-scale storage are a potential risk.
For each mutation supplied, NAM assesses whether there is a risk to the pipeline. If that risk triggers further action, the pipeline manager conducts a field inspection. With NEO's mutation detection, the physical inspection can be carried out in a more focused way. This means an increase in efficiency while ensuring the completeness of risks.
