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

Dynamic aero-optical effects under high-speed flight conditions can severely impair the detection performance of infrared imaging systems, while existing numerical approaches often fail to simultaneously achieve the high accuracy and computational efficiency required for rapid prediction in full-field, multi-spectral scenarios. In this work, a full-chain, end-to-end integration of optical-field dynamics with transient fluid–structure–thermal multiphysics coupling is established, forming a unified aero-optical light-field transmission model spanning spatial, temporal, spectral, and energy dimensions. The multi-dimensional light field is parameterized using low-dimensional, continuously differentiable representations, and a multilayer perceptron integrated with cellular automata (MLPs–CA) parallel ray-tracing framework is developed to efficiently solve coupled transmission across rays, multiphysics fields, optical systems, and sensors, while preserving high accuracy and fidelity. Numerical simulations indicate physical accuracy with point spread function (PSF) structural similarity index (SSIM) ≥ 0.98, together with a computational speedup of 1.29×–9.87× compared to conventional approaches. Analysis under two representative flight conditions further shows that: (1) dynamic aero-optical effects exhibit pronounced sensitivity to temporal, spatial, and spectral variations across different fields of view, with central-field image quality exceeding that of edge fields; and (2) aero-thermal radiation induces temporally, spatially, and spectrally correlated non-uniform image-plane radiance, resulting in average signal-to-noise ratio (SNR) decreases of 33.94% and 40.23% under a 500 K blackbody input. This framework offers a rapid, high-accuracy, and high-fidelity prediction capability for dynamic infrared imaging performance degradation, providing a robust basis for optimizing infrared imaging systems within realistic and complex aero-optical environments.