Spatial resolution of a first-generation gamma ray-computed tomography system

Abstract

First-generation gamma ray–computed tomography systems (FGCTS) provide a cost-effective approach for non-destructive imaging, but are inherently time-intensive, and image quality and spatial resolution strongly depend on acquisition parameters and reconstruction techniques. To balance these trade-offs, this study identified the optimal parameter configuration of the GORBIT FGCTS and evaluated its spatial resolution using analytical image reconstruction techniques. Optimization of parameters was conducted using a single-projection linear scan with varying energy windowing and counting times. Performance was assessed based on photon count statistics, scan duration, and the ratio of mean counts between the spatial resolution (SR) phantom and air regions.  Among the configurations, the parameter without energy windowing, with one-second counting time, C1, was selected as optimal, as it provides sufficient photon counts for image reconstruction and material discrimination with significantly lower total scan duration. Using the optimized parameters, projection data were acquired from a 3D-printed SR phantom with cylindrical features ranging from 10 mm to 3 mm in diameter. Images were reconstructed using backprojection (BP) and filtered backprojection (FBP) with Ram-Lak, Shepp-Logan, Hamming, Hann, and Cosine filters. Quantitative evaluation using Structural Similarity Index Measure (SSIM) and Root Mean Square Error (RMSE) showed that FBP with the Hann filter achieved the best reconstruction performance (SSIM = 0.9712, RMSE = 0.00208). Qualitative and line profile analyses showed that the system resolved features ≥ 4.5 mm, corresponding to a spatial resolution between 4.0 mm and 4.5 mm.