Automated Lightning Flash Detection in Nighttime Visible Satellite Data

The Defense Meteorological Satellite Program (DMSP) Operational Linescan System (OLS) nighttime visible channel allows for the detection of clouds and snow cover only with the availability of moonlight reflection. However, visible light emission from self-contained light sources such as city lights, fires, lightning-illuminated thunderstorms, boats, gas flares, and auroras can be imaged more generally during nighttime hours. This research examines the ability to automate the identification and enhancement of bright steaks associated with lightning flashes and thunderstorm electrification in OLS nighttime visible data in all degrees of lunar illumination. A line-directional filter is applied to the data in order to initially identify these bright streaks and a Red-Green-Blue (RGB) composite product of nighttime visible, infrared, and filtered image output is created to enhance uniquely identifiable lightning streaks while minimizing false detections. This OLS enhanced lightning detection imagery is useful in finding or confirming electrically-active storms anywhere on the globe. It is particularly useful in remote areas for which surface-based observations are not available.