On this page
ClearObject built the Vision Aerial Intelligence Platform for global power company AES on Google Cloud Platform. Its computer vision model cut the number of drone inspection images needing human review by 55% after the first round of training and made wind turbine inspections in the field 10x more efficient.
AES generates and distributes electric power in 15 countries, and windfarms in various geographical locations help them do it as part of their renewable energy generation portfolio.
The Challenge: Automating visual aerial inspections
Operating all the wind turbines in a windfarm is no easy task. Nor is performing visual aerial inspections for maintenance. Even with drones taking visual images, the problem was that humans still had to review every image to identify equipment readiness and maintenance issues. Hundreds of thousands of images, tedious manual reviews — do the math.
The Solution: Vision Aerial Intelligence Platform
The platform is Google Cloud Platform (GCP). ClearObject instrumented a data processing pipeline within the GCP environment, and then developed a computer vision model to reduce the volume of inspection images to be reviewed using machine learning.
The Result
55% reduction in images requiring human review after the initial round of vision model training. As the model continues to learn and improve over time, the image reduction rate is expected to improve to 90%.
10x more efficient wind turbine inspections in the field
What AES had to say…
“Working with Google, we identified ClearObject to be our local GCP partner to help us architect and develop our platform using the latest thinking in cloud and serverless tools available from Google. ClearObject has been a great partner and worked to quickly develop this platform for us.”
–Nicholas Osborn, Global Machine Learning PMO, The AES Corporation
When you understand that every customer is unique and consistently deliver what they ask for, the result can be a whole library of “success stories.”
See more on our Customer Success Stories page.
Frequently asked questions
What problem did AES need to solve?
Drones captured visual images of its wind turbines, but humans still had to review hundreds of thousands of images to identify equipment readiness and maintenance issues.
How much did machine learning reduce manual image review?
Images requiring human review dropped 55% after the initial round of vision model training, and the reduction was expected to improve to 90% as the model kept learning.
What platform did ClearObject use?
Google Cloud Platform. ClearObject built a data processing pipeline in the GCP environment and then developed a computer vision model to reduce the volume of images to review.