Speaker
Description
During the assessment of our civil infrastructure, engineers collect vital information in the form of images, structural drawings, readings (from sensors), and make informed decisions about the state of the structure supported on this information. Deep learning and computer vision have proved powerful tools for automating the information extraction from these data sources and supporting the inspection activities. This presentation will be divided into two parts: (1) Automating the extraction of unbiased and task driven information from more than 100,000 building images and 10,000 bridge images. (2) Using historical data to recommend and update decisions based on changing conditions. Emphasis will be made on considerations to be made in balancing the costs and risks associated with increasing the levels of AI involvement to enable engineers to better manage their resources. This presentation will discuss the methods and share important lessons from these investigations on the power of artificial intelligence to aid the work of the engineer in performing these tasks.
