Situational Information Guidance for Revised Radiological Detection Thresholds

Through the highly popular method of data-fusion, a more intelligent radiological detection system is being researched at the University of Florida. The addition of 3D vision sensors (or depth sensors) to a passive nuclear material detection system in dynamic environments is proving to be beneficial. With 3D vision sensors, nuclear material can now be detected and. The combination of sensors, radiological and depth, have complementary strengths and can enable applications such as tracking behind walls and detecting multiple radiological sources in the same scene. The focus of this paper is to show how 3D vision sensors can provide a priori information such as distance and object speed to improve the detection threshold of a radiological detection system. If a source passed by a radiological detector, the detector response would follow a specific trend over time. With the input from the 3D vision sensor we can fit an equation to the radiological data over time. We compared two fit methods (based on Gaussian and rational equations) to the original detection threshold definition defined by Currie. For scenarios where the count rate of a passing source is discernable from the background count rate, all of the methods performed similarly. However, for instances where the count rate of a passing source is not noticeably above background, the revised detection threshold method limits (fit methods) outperformed the original detection limit.