Machine Vision For Imaging Bubbles: Neutron Detector Technology For Zero-Knowledge Warhead Verification

Abstract

This paper presents new developments in neutron detector technology for zero-knowledge warhead verification. In it, we show that non-electronic neutron detectors, conventionally known as bubble detectors, utilizing superheated halocarbon emulsions can be used successfully as a form of detector physics; such verification tools are able to protect sensitive nuclear information more so than current complex electronic systems, which hide such classified information behind engineered information barriers. Additionally, we build upon the current bubble counting technique by suggesting a completely new and improved counting method: we propose using multiple images of each bubble detector taken at many angles as the detector rotates to vote on bubble location, and thus provide more accurate measurement of bubble counts at higher neutron fluence. We begin by exploring the concept of a zero-knowledge proof and applying it in the context of nuclear arms control. We then break down the physics of the bubble detectors in the context of radiation dosimetry, and examine fundamental concepts of neutron interactions with matter. Finally, we introduce the set-up of an experiment currently in place at the Princeton Plasma Physics Laboratory to test and develop the bubble detectors, present the analytical process we developed to analyze the bubble detector images, and ultimately provide an accurate measurement of bubble count at high neutron fluence.