Overlapped materials decomposition in high-energy dual-energy X-ray system

α-Curve method has been widely used for material discrimination in high energy container inspection system. However, this method hypothesizes that the object to be detected is pure material, which provides inaccurate atomic number and thickness in the case of overlapped objects. A surface equation theory is used in medical domain for material decomposition which can provide accurate bone mineral density. The goal of our research is to examine the feasibility and accuracy of this theory in container inspection system which uses high energy dual X ray. The thickness of each material can be obtained in the case of overlapped material discrimination if the atomic numbers are given. The experiments were carried out on a container inspection system for alpha curve drawing and conic surface calibration. C, Al and Fe were used as basis materials. The coefficients of calibrated conic surface are used to obtain the mass thickness of each basis material. Alpha curve is used to reconstruct the dual energy data of each material. Then the distribution of each material in the image can be obtained using this algorithm. We quantitatively analyze the accuracy and noise in the case of different material combinations and thickness, then discuss the results of original discrimination method and this theory.