Improving image reconstruction algorithm for rotating modulation collimators using a variance stabilizing transform

Abstract

A rotating modulation collimator (RMC) is an indirect imaging technique that has proven useful for remote radiation source detection. While it was initially invented for detecting sources in a far field, a recent development by Kowash has shown the feasibility of the RMC for detecting mid-range sources. However, their image reconstruction algorithm often produces spurious source estimates in pixels where no source exists. In this paper, we propose to improve the RMC image quality using a variance stabilizing transform. The transform reduces the inhomogeneous Poisson noise in the RMC data. In our simulation study, the proposed algorithm could reconstruct RMC images without misleading artifacts.