[Paper Review] Gamma-Ray Imaging with the Coded Mask IBIS Telescope

A. Goldwurm, P. Goldoni+7 more|arXiv (Cornell University)|Feb 22, 2001

Gamma-ray bursts and supernovae3 citations

TL;DR

This paper presents the coded mask imaging methodology for the IBIS telescope on the INTEGRAL mission, detailing how coded aperture techniques enable high-sensitivity soft gamma-ray imaging (20 keV–10 MeV) using iterative deconvolution and correlation-based reconstruction. Key results show that IBIS/ISGRI can locate a 30σ point source with sub-arcminute accuracy (±30″), limited primarily by attitude knowledge rather than instrumental resolution.

ABSTRACT

The IBIS telescope onboard INTEGRAL, the ESA gamma-ray space mission to be launched in 2002, is a soft gamma-ray (20 keV - 10 MeV) device based on a coded aperture imaging system. We describe here basic concepts of coded masks, the imaging system of the IBIS telescope, and the standard data analysis procedures to reconstruct sky images. This analysis includes, for both the low-energy detector layer (ISGRI) and the high energy layer (PICSIT), iterative procedures which decode recorded shadowgrams, search for and locate sources, clean for secondary lobes, and then rotate and compose sky images. These procedures will be implemented in the Quick Look and Standard Analysis of the INTEGRAL Science Data Center (ISDC) as IBIS Instrument Specific Software.

Motivation & Objective

• To develop and validate a robust, iterative imaging pipeline for the IBIS coded mask telescope on the INTEGRAL mission.
• To enable high-sensitivity, wide-field soft gamma-ray imaging in the 20 keV–10 MeV range where conventional focusing is impractical.
• To minimize imaging artifacts and background contamination through optimized mask patterns and iterative source subtraction.
• To quantify the system's angular resolution and source localization accuracy under realistic observational conditions.
• To implement the method in the INTEGRAL Science Data Center (ISDC) for Quick Look and Standard Analysis pipelines.

Proposed method

• Uses a coded aperture mask with a MURA (Modified Uniformly Redundant Array) pattern to cast unique, modulated shadows of celestial sources onto the detector.
• Employs cross-correlation with a decoding array G = 2M - 1 to reconstruct sky images, where M is the mask pattern, ensuring sidelobe-free response for fully coded sources.
• Applies iterative image reconstruction: source detection, PSF fitting, source model subtraction, image rotation, and variance-weighted combination.
• Corrects for detector non-uniformity and background using flat-fielding and statistical normalization techniques.
• Uses chi-square fitting of deconvolved image sectors with the System Point Spread Function (SPSF) to estimate source position and intensity errors.
• Simulates observations to validate performance, varying source S/N and mask-to-pixel size ratios (R = 2.43 for ISGRI).

Experimental results

Research questions

• RQ1 How can coded aperture imaging achieve high angular resolution and sensitivity in the soft gamma-ray band (20 keV–10 MeV) with minimal background degradation?
• RQ2 What is the optimal mask pattern that minimizes secondary lobes and ensures position-invariant imaging response?
• RQ3 How does the signal-to-noise ratio (S/N) and mask-to-pixel size ratio (R) affect source localization accuracy in real observations?
• RQ4 What iterative procedures are necessary to clean residual artifacts from source side lobes and improve image fidelity?
• RQ5 To what extent can the IBIS/ISGRI system resolve point sources with high S/N, and what are the limits imposed by instrumental and pointing errors?

Key findings

• The IBIS/ISGRI telescope achieves a full-width at half-maximum (FWHM) angular resolution of 12.0′ for the ISGRI layer, with a pixel size of 5.0′.
• For a 30σ point source, the system can localize the source to within ±30″, assuming no systematic errors beyond attitude knowledge.
• The absolute error in attitude reconstruction for INTEGRAL is expected to be less than 20″, which is the dominant error source for source localization.
• The system’s imaging efficiency for ISGRI ranges from 0.86 to 0.81, depending on source position, due to detector and mask geometry.
• Theoretical and simulated position errors scale as (S/N)⁻¹, with 90% confidence level errors matching predicted limits for R = 2.43.
• The system achieves constant sensitivity and optimal image quality within the fully coded field of view (FCFOV) of 8.3° × 8.6°, while sensitivity degrades in the partially coded FOV.