Our AI image recognition has been trained with over 50,000 images and videos and is able to recognize bolides (meteors) from all image and video files. The AI analyzes the image or video file and compiles a separate image file upon sighting.
During the analysis process, the AI automatically calculates object properties of the bolide/meteor based on different metadata and algorithms and, in the case of identical objects (cases), also the possible fall area.
Once the fall zone is calculated, we use the same AI image recognition on our finder drone, but this time trained on meteroites. Here the AI creates a grid of the fall area and creates an automatic flight path for the drone. The camera and image recognition now reports objects that come close to a meteorite and transmits the coordinates and screenshots of the possible meteorite hits.
The goal of the ASO22 project is to use AI (machine learning) to enable fast and reliable detection of meteors (bolides) in the night sky, calculation of fall areas and detection of meteorites.
ASO22 starts at July 2021.
Meteor & Meteorite Researcher & Developer
@ Neorc, Hoovi
Front & Backend Developer & Data Science @ Hoovi
KI Developer & Data Science @ Hoovi