The Machine Vision Radiation Detection System

The emerging threats created by a global expansion of nuclear technologies and terrorism demand improved nuclear materials detection systems to aid in nuclear security and nonproliferation. This project develops the idea of using machine vision combined with a large-area gamma-ray imager to improve sensitivity to threats and their rapid localization in a crowded environment (e.g., subway stations, airports, and bridges). We have achieved our first coded-mask images with a 1 m² array of 100 NaI(Tl) detectors. In addition, two video cameras have been used in stereo to create a three-dimensional map of points in front of the array, and image segmentation is being implemented to distinguish and track individual objects in the field of view. We are currently gearing up to simultaneously perform real-time gamma-ray imaging and object tracking so that we can eventually merge the two data streams and achieve the expected increase in sensitivity of this method.