Helios: real-time adaptive weather surveillance system using integrated camera networks and enhanced geospatial processing algorithms

The need is stronger than ever to be able to autonomously maintain persistent surveillance of weather events over a selected region of interest on a continuing basis, as well as automated method of tasking visual assets on weather activity. Real-time confirmation of weather events at a hyper-local level is a new and important aspect of forecasting. The HELIOS project looks to fill the void in weather data dissemination of real-time coordinated ground-based video around confirming a weather event by use of an integrated network of cameras and enhanced geospatial processing algorithms. Historically, weather observation networks have focused on the atmosphere and ignored the area below the horizon where people live and drive. Thus, a host of businesses, government agencies and the public must rely on field observations, social media or weather data that is primarily tailored to aviation for daily decision making. For example, transportation and electric utilities can take mitigating action on impending severe weather; severe weather diagnosis products are useful beyond just weather forecasters. This paper investigates a real-time adaptive surveillance system that provides live, targeted weather event video/still imagery specific to severe weather events, as well as additional algorithmic capabilities to detect specific weather phenomena (rain, hail, tornado funnels) from ground video. Additional services include archival, time-lapse products and streaming of weather phenomena to customers using existing proprietary systems. NWS rebroadcast data is used to provide storm location, severity, bearing, etc. of weather events. The HELIOS system re-fixes or selects appropriate viewing existing camera infrastructure (traffic, etc.) on the progressing storm from multiple vantage points. This data is feed 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. We present a real-time discussion and demonstration of the proposed technique with video collected from our Beta persistent surveillance product. Our approach maintains compatibility with existing GIS databases and provides an integrated solution for multi-source video fusion of localized weather events.