A correction method for dual energy liquid CT image reconstruction with metallic containers.

J Xray Sci Technol

Department of Engineering Physics, Tsinghua University, Ministry of Education, Beijing, China.

Published: January 2013

With its capability of material discrimination, dual energy computed tomography (DECT) is widely used in security inspection for the purpose of detecting contraband. DECT provides effective atomic number image and electron density image in addition to traditional attenuation images. In dual energy liquid inspection system, the presence of metallic containers will cause partial volume effect (PVE) that leads to severe deviation in effective atomic number image. Usually, the deviation is too large for a reliable material discrimination and may cause false results. In this paper, a projection splitting method is proposed to combat the PVE. This method is based on the assumption that a prior projection of the empty container is available and photoelectric and Compton coefficient integrals can be calculated via dual energy decomposition. Each integral is split into two parts by subtracting the integral of the empty container from the total integral. The subtraction removes the integral part contributed by the container, thus discarding the error source created by PVE that appears on the boundary of the sinogram. Images are reconstructed in which only the interior liquid area remains. Experiments are performed in a real liquid inspection system to demonstrate the effectiveness of this method. Accuracy of the reconstructed effective atomic number is greatly improved with this method, which helps a lot in determining the type of the object.

Download full-text PDF

Source
http://dx.doi.org/10.3233/XST-2012-0339DOI Listing

Publication Analysis

Top Keywords

dual energy
16
effective atomic
12
atomic number
12
energy liquid
8
metallic containers
8
material discrimination
8
number image
8
liquid inspection
8
inspection system
8
empty container
8

Similar Publications

Want AI Summaries of new PubMed Abstracts delivered to your In-box?

Enter search terms and have AI summaries delivered each week - change queries or unsubscribe any time!