Joint Source Activity And Motion Reconstruction In A Coded-Aperture Gamma Camera For Home-Land Security Applications

Yan Xia,Qing Ye,Rui Wang,Peng Fan,Yaqiang Liu,Tianyu Ma
DOI: https://doi.org/10.1109/nssmic.2018.8824684
2018-01-01
Abstract:A coded-aperture gamma camera has been developed for home-land security purpose. In passenger inspection applications, the image suffers from severe artifact when a weak source moves across the FOV. In this study, we propose a novel image reconstruction algorithm that jointly estimate the activity and motion of a point source. We implement an alternate activity-motion update strategy. An image-domain sparsity prior is also introduced into the MAP reconstruction framework. The efficacy of this algorithm was validated through numerical simulations. Using the system matrix from the real system, point source projection data were generated at three different count level (1M, 100 and 20 counts/ frame) with two pre-defined source moving paths. With conventional frame-by-frame MLEM reconstruction, correct single-point image is achieved at 1M count/frame level. Noticeable artifact appears at 100 counts/frame level. In 20 counts/frame, the MLEM algorithm fails to generate reasonable image. In joint reconstruction, in all three cases the reconstructed image converge to a single point. The position accuracy is feasible for localizing suspect passenger even at 20 count/frame level case. We conclude that the proposed algorithm is very promising for improving the passenger inspection efficiency.
What problem does this paper attempt to address?