Big Data takes to the skies with aerial monitoring

By: Vladimir Nedashkovsky

For industries like utilities, energy, forestry and mining with extensive assets on the ground, keeping track of maintenance and other monitoring activities is time consuming and very costly. Power cables can extend for hundreds of kilometers, often through difficult terrain. Forestry plantations can cover hundreds of thousands of hectares. Gas pipelines often run through inaccessible and wild locations.

In the absence of any realistic alternatives, maintenance ground crews are deployed to physically monitor and inspect these assets. But in many cases, they are only able to cover a small proportion of assets’ full extent. For example, only 1-2% of a forestry planation of a million hectares can be feasibly inspected in any single year. And that means problems will go unseen.

Aerial photography has in some cases helped to create more effective situational awareness for asset health. But even here the analysis of resulting imagery is a painstaking and manually-intensive process. Satellite imagery, while increasingly affordable, is not available in the resolution required to provide a sufficiently detailed picture of many key assets on the ground.

However, by combining existing technologies – unmanned aircraft and computer vision, big data and predictive analytics – it is now becoming feasible to achieve detailed imagery and automated object detection and analysis that can dramatically improve the speed, extent and accuracy of asset monitoring and maintenance. Integration of these technologies can help businesses charged with maintaining extensive physical assets – a particular concern for many businesses in the Nordics – to achieve more effective results and predictive analysis of potential problems at a considerably lower cost than the manual methods they use today.

Unmanned aerial or underwater, remotely operated vehicles –drones, UAVs, ROVs – can be equipped with a range of digital imaging and spectral sensors to capture data over extensive physical assets. For industries such as energy, utilities, agriculture and forestry, the potential for better results, more efficient use of maintenance resources and far lower costs, is significant. Whereas today the visual data from an inspection of 25 hectares of forestry will take two operators four days to check through manually, using big data analytics can create more accurate (and therefore useful) results in a matter of hours. It’s the same story under the water. Sub-sea engineers typically spend up to six months analysing six days of subsea video footage to complete an inspection report. It takes up to a week for a maintenance team to inspect 300-500 power lines and report the defects they find. Using video analytics can complete these tasks within 24 hours.

For power companies operating extensive lines, vegetation control can take up to 40% of the maintenance budget, and many months, to make sure that the channels through which power lines run are clear of obstruction and that there is easy access to assets that require repairs and maintenance. In addition, spotting unauthorised access and use of land, detecting broken isolators, sagging power lines and tower tilting all also need to be monitored. Yet using a combination of UAVs, computer vision and big data analytics can see the same tasks achieved in a single day – from image capture of an entire powerline through to initial analysis and results.

The technology can also be used to improve the efficiency of inspecting offshore oil and gas production assets. Today, shutting down a platform before an inspection is carried out is extremely costly. Lost production can run into millions of dollars. Carrying out a pre-inspection using unmanned aerial vehicles can help cut the duration of shut down to identify key targets for maintenance and make considerable savings.

And as imaging and sensor technologies evolve, there are already further innovations coming online. Multi/hyper-spectral ‘cameras’, for example, can detect indicators such as chlorophyll or CO2 levels, to reveal otherwise hidden aspects of vegetation health across a forestry plantation. A UV camera can detect corona discharge, thermal cameras can be used to spot various types of leaks such as oil, heat or steam from asset installations that could indicate underlying problems.

For companies in the Nordics that need to maintain physical assets covering large amounts of often challenging terrain, the innovative combination of UAVs, computer vision and big data analytics could prove a game-changer. To maximize the health of their assets on the ground they may need to start looking to the sky.