A reliable and efficient distribution grid requires regular maintenance of its electrical and physical infrastructures. In person, individual physical inspections for preventive maintenance require extensive resource management and are often not feasible for large distribution grid operators. Therefore, an AI-based grid infrastructure defect detection tool can provide value through efficiency and scalability.

“Image AI” is a computer vision-based application developed by the Advanced Analytics group at Exelon that detects distribution grid infrastructure defects (poles, cross arm, fuse, pin arm, etc.). The application provides actionable intelligence by analyzing raw grid infrastructure images captured through UAVs (Unmanned Aerial Vehicles), satellites, or through other resources.