Using Laser Scanners as Smart Sensors to Detect and Categorize Entry/Exit Events into/from a Plutonium Store

The DG JRC and the DG ENER have designed and installed detection, measuring and containment & surveillance equipment pieces in a plutonium oxide store in a reprocessing plant to perform the necessary safeguards activities with limited manpower resources while allowing the Operator to access the store even in absence of the inspectors. Among the equipment pieces, there is the Double Laser Curtain (DLC), which consists of two laser scanners installed over the entrance door to detect any entry or exit event into or from the store. The DLC is operational from the beginning of the project to collect the data which is used to draw “profiles” which are characteristic of the “most common objects” which go through the DLC beams. The analysis of the profiles allows to distinct entry/exit of people, of a person pushing a trolley for the transfer of PuO2 cans, of a person pushing an “alpha in air” monitoring system, the passage of a forklift truck equipped with a heavily shielded block used for the transfer of the cans inside the store, or the passage of a forklift truck with a stillage for the export of PuO2 cans. The DLC is only a partial contribution to an integrated solution and not a self-standing safeguards measure by itself, but combined with other data it performs in a very promising way. In fact, whereas the overall project was described in previous papers, this paper describes in more detail the DLC equipment, its installation, the data collection and storage, the evaluation of the normalized profiles to perform the categorization of the entry and exit events to facilitate the review, with a dedicated application, of the (only) pertinent images by the inspectors to make sure, with the other information collected, that no diversion of nuclear material took place and that the procedures are respected by the Operator. To evaluate the performances of the classification algorithms, the results returned will be compared with the ones obtained by “human eye” classification of several thousands of profiles collected during two years of operation of the DLC which could also find applications in other facilities.