AVIS (Autonomous Video Inspection System) is an innovative highway-rail grade crossing inspection system that will use deploy Artificial Intelligence based models with efficiency and accuracy. AVIS is an end-to-end solution that will evaluate forward-facing video to detect the presence and condition of key features that help improve the safety of pedestrians and vehicles at highway-rail crossings.While crossing inspections are currently performed manually, ARIS will automate this process and provide a framework for users to understand which crossings have been inspected and see the results alongside other relevant data such as geometry data and drone imagery. The machine learning models will be developed so that they can re-learn as more footage is processed helping to improve accuracy over time. Further, AVIS will use clustered-computing and be able to ignore sections of video that do not contain crossings to help improve the delivery time of inspection reports to key stakeholders.
