The Direction of Automated Visual Inspection for Industrial Assets

/

Percepto, an autonomous inspection and monitoring solution provider, recently surveyed 200 senior-level oil and gas professionals involved in asset inspections to gauge the level of automation and data analysis. The survey reveals that adoption is increasing but is behind where many expect it to be based on the available technology—with most automated programs only at the largest producers. 

Survey respondents indicated that their main reasons for conducting regular inspections is safety. Yet most are still relying on humans performing manual processes to inspect assets.

“Drones, robotics and AI allow for automated visual infrastructure inspections at a pace, scale and accuracy level that humans alone cannot deliver using traditional visual inspection methods, particularly for hazardous, hard-to-reach assets,” said Percepto Co-founder and CEO Dor Abuhasira. “The heavy reliance on labor intensive, manual methods suggests a lack of awareness of how AI can be utilized to ensure the highest levels of safety, productivity, asset reliability and adherence to regulations, representing a vast untapped technological resource.”

The lack of adoption is surprising. 81% of respondents indicated they are not currently using drones for visual inspection and 83% are not using artificial intelligence (AI). These respondents do plan to integrate drone and AI technologies soon, but run into budget and buy-in challenges. Respondents, however, aren’t as bought-in on robots for visual inspection, with only 29% planning to use them, limiting visuals to aerial drone imaging. 

Most companies are collecting visual data—photos and videos of assets—but few, only 12%, are using AI or machine learning (ML) to analyze the imagery for insights. Analysis is being completed manually or not at all. 

While drone and AI adoption are increasing, end-to-end automation is still far from a reality for most oil and gas companies. 94% of respondents are using some level of automation, but only 30% say they’re automating both the data collection and analysis processes. 

Because of reliance on manual inspections, assets are not being inspected as frequently as would be ideal. Assets that are hardest to inspect—tanks, pipelines, power lines, electric poles—are inspected the least and yet can have disastrous outcomes if something goes wrong. 

All of the respondents still rely on manned visual inspection and many are still using in-house resources and analysis tools rather than transitioning to best-of-breed industry innovations. 

The full report is available on Percepto’s website. 

Solution providers like Percepto and others are offering asset owners an easier entry into automated inspection and analysis programs. This transition, like inspections themselves, won’t be automatic.