Analysis Automation for High Explosive Breakout Symmetry

Sean Tronsen, Elizabeth Francois, Christina Scovel, Nathan DeBardeleben

Abstract: In this work, we present an automated approach for analyzing hemispherical breakout symmetry in high explosive detonation experiments. Traditionally, the analysis of streak camera records from these tests required extensive human intervention to ensure accurate measurements, thereby slowing the rate at which additional tests could be fired. Our developed procedures leverage a combination of classic vision algorithms, probabilistic algorithms like RANSAC, and machine learning techniques, reducing the analysis time from several hours to several minutes. The software was designed with the intent to run on non-exotic consumer hardware, thus enabling researchers to obtain results in the field using standard laptops within minutes.

External IDs:dblp:conf/isvc/TronsenFSD24