An AI-based visual recognition system aims to increase the performance of the existing mechanism that is used to monitor and defend the controlled airspace over Washington, D.C. The Defense Innovation Unit (DIU) — a Department of Defense organization focused exclusively on fielding and scaling commercial technology across the U.S. military at commercial speeds — is leading development of the technology. Los Angeles-based visual tracking software developer Teleidoscope is the recipient of a production contract for up to $100 million for the AI laser system. According to the DIU, the undertaking will lead to technology with broad national defense applicability for both defense against asymmetric and near-peer threats like unmanned aerial systems and cruise missiles. The to-be-developed systems leverage commercial market advancements in cameras, computer vision, augmented reality, and machine learning to improve the air defense operator’s ability to detect, identify, and track aircraft. The Teleidoscope team demonstrates upgraded electro optical/infrared cameras with enhanced capabilities to improve airspace awareness. Courtesy of Defense Innovation Unit. The airspace around Washington, D.C., is under the National Capital Region-Integrated Air Defense System (NCR-IADS), which was created after Sept. 11, 2001, to provide low-altitude, low-airspeed detection capabilities to defend against airborne threats to the NCR. Now, the NCR-IADS modernization effort intends to upgrade system cameras and eye-safe lasers used for tracking and visually warning aircraft in violation of the special flight rules within the NCR. The updates improve air defense operators’ ability to identify aircraft and aim warning lasers at much farther ranges. The auto-tracking capabilities of the system are applicable to full-motion-video feeds, regardless of the domain, opening the door to augment remotely piloted aircraft video feed tracking capabilities, the DIU said. The prototype has the potential to run on any edge device or cloud-provided full-motion video feed.