A Rapid Damage Assessment System Using Machine Vision to Detect Damage to Utility Grid Infrastructure

Power outages from natural disasters not only result in billions of dollars of losses, associated health and safety impacts can result in significant loss of life. After a disaster, damage assessments of electrical grid infrastructure are critical for developing a strategy for power restoration including the optimal allocation of resources. Despite improvements, the damage assessment process is still a slow, manual process, which impacts the speed of recovery. In the Phase I effort, Oceanit developed and demonstrated an automated assessment system with real- time on board image processing to identify damaged grid infrastructure from a drone. Oceanit’s ePoP technology would be used to provide wireless communications and power and Oceanit’s MERCI technology brings command and control capabilities. In the Phase I effort, Oceanit used its VIPA (Versatile Image Processing Architecture) machine vision technology to identify grid infrastructure and commercial off-the-shelf machine learning frameworks to categorize damage real-time. Oceanit’s MERCI (Mobile Emergency Response and Command Interface) software was customized to interface with the machine learning framework so that damage assessment data would be consolidated in a single data source for actionable intelligence. Oceanit’s ePoP, pop up power and communications system was configured with an external gateway so that a satellite Internet device could be used so the system could send damage information to a centralized Emergency Operations Center (EOC) even if there is no power or cellular communications in the field. For the Phase II effort, the object identification and object classification algorithms will be optimized and integrated into the MERCI and ePoP system, which will be packaged into a field deployable system that we are calling MERCI-UI (Utility Inspections). MERCI-UI will be tested and evaluated by strategic partners as pricing and go-to-market strategies developed for targeted markets. The effort will also focus on industry engagement opportunities including events and demonstrations to position MERCI-UI for commercialization. Commercial Applications and Other Benefits: In addition to improving the damage assessment process for utility inspections, MERCI-UI could be used for day to day inspections as well as inspections in other vertical markets including the oil and gas, agriculture, and the transportation industries.