A fitness-guided adaptive genetic algorithm for near-field neutron coded imaging reconstruction

Area of Science:

• Nuclear Fusion Diagnostics
• Computational Imaging
• Applied Mathematics

Background:

• Neutron coded imaging is crucial for diagnosing inertial confinement fusion (ICF) reaction zones.
• High-fidelity reconstruction in near-field conditions presents an ill-posed inverse problem.
• Traditional methods often yield binary reconstructions, limiting diagnostic detail.

Purpose of the Study:

• To develop an advanced algorithm for high-precision near-field neutron coded imaging reconstruction.
• To address the challenges of ill-posed inverse problems in ICF diagnostics.
• To improve upon existing reconstruction techniques for neutron sources.

Main Methods:

• Introduction of a novel Fitness-Guided Adaptive Genetic Algorithm (FGAGA).
• FGAGA utilizes adaptive selection, hybrid crossover, and adaptive mutation operations.
• Pixel classification and specialized mutation rules enhance convergence and diversity.
• An Isolated Pixel Treatment Operation minimizes reconstruction artifacts.

Main Results:

• FGAGA achieves high-precision continuous grayscale reconstruction of near-field neutron sources.
• The algorithm performs effectively even under low neutron yield conditions.
• Demonstrated superior performance over traditional heuristic and deterministic algorithms like SART with TV.
• FGAGA achieved a 2.27 dB gain in PSNR and 0.12 increase in SSIM compared to SART with TV.

Conclusions:

• FGAGA offers a significant advancement in near-field neutron coded imaging reconstruction.
• The algorithm overcomes limitations of binary reconstruction, providing richer diagnostic information.
• FGAGA demonstrates robust performance and accuracy for ICF applications.