Towards the use of Scene Data Fusion for Radiological Decontamination & Decommissioning Applications

Abstract

Characterizing radiological contamination is crucial before decommissioning begins at nuclear facilities. Detecting, localizing, identifying, and quantifying these sources is required to develop an optimal decommissioning strategy while ensuring worker safety. The current approach utilizes a combination of historic records, estimations using analytical modeling with known operational information, and onsite surveying. Onsite measurements are time-consuming due to complex geometries, unknown/shielded sources, and by the presence of non-radiological hazards. Free-moving 3-D Scene Data Fusion (SDF) can improve characterization efforts through fusion of radiological data with contextual imagery and position sensors. This is achieved by integrating a radiation detector with contextual sensors that provide real-time tracking of the system pose as it moves through the environment. Together with advanced algorithms, gamma-ray image reconstruction is performed in real-time resulting in a 3D model of the environment with radiological sources attributed to objects within the scene. We describe the SDF approach and systems of various form factors developed by Lawrence Berkeley National Laboratory (LBNL) appropriate for nuclear power plant/laboratory Decontamination and Decommissioning (D&D) activities and environmental remediation applications. Using these systems, we demonstrate SDF to characterize radiological sources within a 3D scene with high accuracy (source localization, nuclear material holdup, Fukushima decontamination verification).