Lightning is one of the more mysterious and interesting phenomena on the planet. Very powerful, but each strike on average has enough power to power an incandescent bulb for an hour. The exact mechanism that initiates the lightning strike is not yet well understood. But it happens 45 times per second somewhere on the planet. Although we can’t get a deeper scientific appreciation of lightning anytime soon, but we can capture it in different photos thanks to this project that uses machine learning to get the best lightning frames .
The creator of the project, [Liam], it was created as a tool for stormchasers and photographers so they could film for hours and no longer have to replay their footage manually to capture frames with flashes. The project borrows from a similar project, but it adds Python 3 capability and runs on a small netbook for faster deployment in the field. It uses OpenCV for object recognition, uses video files as source data, and has different modes to identify different types of lightning.
The software is free and open source, and releases are supported for Windows and Linux. Currently, [Liam] able to capture all kinds of electrical atmospheric phenomena with it including lightning, red sprites, and elves. We don’t see very many projects involving lightning here, in part because people can only produce a fraction of the voltage potential needed for typical lightning.