A large utility client needed to make their asset inspections even faster, cheaper, reliable, accountable, and auditable. To accomplish higher volume inpections, this required a new approach (remote inspectors) since they were limited by the workforce in rural areas.
Nexient supported the roll out an MVP for drone image inspections in a first for client 10-week period. Providing 3 full stack developers, the team developed a front-end application which allows inspectors to review remote inspection images from multiple circuits of steel structures, wood and other tubular steel poles. The app includes GraphQL server running on docker and deployed to Kubernetes cluster on AWS.
"AI & Machine Learning models were incorporated for analyzing images and identifying subtle changes or damages to equipment."
by reducing the number of apps inspectors use.
and reduced risk of accidental deletion of crucial images.
by decreasing time from about 80 to 52 minutes per structure.
The final, award winning product greatly increases the inspections by reducing the number of apps inspectors use, along with reducing the chances of multiple inspectors working on the same structure. The improved image quality and reduced risk of accidental deletion of crucial images enables a machine learning roadmap for the future. Overall, inspection time decreased from about 80 to 52 minutes per structure – by analyzing images and identifying approach paths more safely, quickly, and accurately.