NASA has long used liquid hydrogen as a fuel and plans to continue using it in association with their advanced nuclear thermal propulsion technology. Hydrogen fire detection is critical for rocket propulsion safety and maintenance. A significant fire at a rocket test or launch facility could be catastrophic to infrastructure or even worse, to human life. Detection monitoring is problematic as hydrogen flames can be nearly invisible during the day. Non-imaging, Non-visible fire detection technology has limited range and can suffer from false alarms from sources outside the region of interest. Low-cost visible imagers, commonly used for wide-scale routine surveillance, have limited utility detecting hydrogen fires. Although it has been known for decades that multispectral imaging outside the visible range can be used to detect fires with low false alarm rates, the price of such systems and the lack of processing algorithms and ability to implement them in real-time has largely prohibited their use. During this project we will develop a low-cost imaging capability that fuses data collected from sensors operating in the (1) solar blind ultra-violet, (2) thermal infrared, (3) mid-wave infrared, and (4) visible spectrum, using advanced spectral, spatial and temporal processing techniques optimized to detect and generate alerts associated with hydrogen fires in real-time. This multi-sensor, multi-processing approach will enable us to automate flame detection with extremely low false alarm rates. This multisensory imaging research could also support NASA's important cool flame microgravity research occurring on the International Space Station.
