Inline multi-material identification via dual energy radiographic measurements

In the waste recycling industry, material separation is of key importance. In this paper, we present a new material identification method, based on dual-energy X-ray radiographic images, developed in the context of waste recycling industry. The algorithm is based on a dual energy technique and allows to estimate the effective atomic number of the sample under investigation. A projection simulator with high accuracy that allows to simulate realistic measurements of a wide range of materials is used. The resulting virtual measurements are then used in an identification tool, which achieves the identification by comparing the virtual and actual measurements. Via this procedure, an extensive range of materials and thicknesses can be analyzed and identified. Finally, an optimization scheme has been developed, which allows the selection of an ideal setting for the scanner, in order to optimize the identification process.