Seminarium dr. Bahadır Saygı
Z przyjemnością zapraszamy na nadchodzące seminarium SLCJ, które odbędzie się w środę,
25 marca 2026, o godz. 12:15 w bibliotece.
Dr Bahadır Saygı z Ankara University wygłosi wykład:
“A Photopeak-Bayesian Isotope Identifier for Radiation Portal Monitors
Using a Medium-Resolution CeBr₃ Detector”
Abstrakt:
Radioisotope Identification Devices (RID) play an essential role in
applied nuclear physics, particularly in nuclear security and safeguards,
where the rapid and reliable characterization of radioactive materials is
crucial. In real-world field conditions, however, identifying
radionuclides in real time—without relying on expert spectral
interpretation—remains challenging. Peak overlap, counting statistics, and
complex background structures often complicate the analysis of gamma-ray
spectra.
In this talk, I will describe the development and experimental validation
of an automated real-time isotope identification algorithm designed to
address these challenges. The approach combines second-derivative-based
peak enhancement with a Bayesian, photopeak-centered decision framework.
Peak detection is carried out using a Mariscotti-type method, allowing
statistically significant gamma-ray photopeaks to be identified in a
robust manner. Isotope attribution is then performed through posterior
probability estimation within a Bayesian inference scheme, taking into
account characteristic gamma emission energies and detector response
effects.
The algorithm has been tested using gamma-ray spectra measured with a
medium energy-resolution scintillation detector. The validation dataset
includes point sources as well as more complex spectra obtained from low
enriched uranium certified reference materials and naturally occurring
uranium–thorium ore samples, representing realistic mixed radiation
fields.
The results show that combining derivative-based peak detection with
Bayesian photopeak classification offers a reliable and computationally
efficient strategy for real-time radionuclide identification. This
framework demonstrates strong potential for implementation in
next-generation, field-deployable gamma spectrometry systems operating
under complex and heterogeneous radiation conditions.
This work has been partially supported by the TUBİTAK 1005 Project
(Project Code: 124F446), titled “Development of an Isotope Identifier
Device (RID) Using a Photopeak-Based Bayesian Statistics Method with
a Medium Energy Resolution CeBr3 Scintillation Detector Integrated
into a Radiation Portal Monitor (RPM).“