Accurate and efficient MLPs-CA-based full-chain dynamic aero-optics modeling for infrared imaging prediction

Area of Science:

• Aerospace Engineering
• Optical Physics
• Computational Fluid Dynamics

Background:

• High-speed flight dynamic aero-optical effects degrade infrared imaging system performance.
• Existing numerical methods lack the accuracy and efficiency for multi-spectral, full-field predictions.

Purpose of the Study:

• To develop a unified, high-fidelity model for predicting infrared imaging performance degradation under dynamic aero-optical conditions.
• To enhance computational efficiency for rapid, real-time analysis.

Main Methods:

• Integrated optical-field dynamics with fluid-structure-thermal multiphysics coupling.
• Developed a multilayer perceptron integrated with cellular automata (MLPs-CA) parallel ray-tracing framework.
• Parameterized multi-dimensional light fields using low-dimensional, continuously differentiable representations.

Main Results:

• Achieved high physical accuracy (PSF SSIM ≥ 0.98) and significant computational speedup (1.29×-9.87×).
• Demonstrated sensitivity of aero-optical effects to temporal, spatial, and spectral variations.
• Quantified signal-to-noise ratio (SNR) decreases due to aero-thermal radiation (avg. 33.94%-40.23%).

Conclusions:

• The developed framework provides rapid, accurate, and high-fidelity predictions for dynamic infrared imaging performance.
• Offers a robust basis for optimizing infrared imaging systems in complex aero-optical environments.