HELIOS: A Real-time Adaptive Weather Surveillance System

The need is stronger than ever for broadcast media to autonomously show their viewers weather events over regions of interest on a continuing basis, as well as automating methods of focusing visual assets on weather activity. This ability would provide a medium for broadcasters to disseminate life saving information to their viewers and the public in general. The HELIOS system fills the void for “eyes-on” ground-based weather data verification of real-time weather events. Using a coordinated network of existing ground-based cameras and enhanced image processing algorithms, the HELIOS system provides a visual asset not currently available to help viewers see and respond to severe weather events.

Historically, broadcast meteorologists have had to rely on weather data primarily tailored to aviation. The need for real-time visual verification at the local level that can provide actionable weather information is paramount. The HELIOS system demonstrates a real-time adaptive surveillance system that provides live, targeted ground-based weather event video of severe weather events, as well as additional algorithmic capabilities to detect specific weather phenomena. Other capabilities include archival and streaming of weather phenomena to customers using existing proprietary systems.

NWS EMWIN rebroadcast data is used to provide storm location, severity, bearing, etc. of a weather event. The HELIOS system selects appropriate existing camera infrastructure (traffic, etc.) to view the progressing storm from multiple vantage points. This data is fed to a video archiving, exploitation and dissemination system and live/archived video products are produced real time. The fusion of coordinated video-based camera networks is provided via the system, allowing the user to select the video feed of interest from the available choices.

We present the results from our prototype system that demonstrates the feasibility of the technique with current camera networks in the Washington, DC area. Our approach maintains compatibility with existing GIS databases and provides an integrated solution for multi-source video fusion of localized weather events.