Enhancing infrared imaging robustness against laser-induced damage: a wavefront coding and Mamba-UNET approach

The high-sensitivity nature of infrared imaging systems makes them highly susceptible to pixel saturation and even detector damage when exposed to unexpected laser radiation, severely compromising imaging reliability. To address the challenge of balancing laser protection and imaging quality in such systems, we employ wavefront coding technology to diffuse the incident laser energy, effectively reducing the energy density on the detector's target area and thereby enhancing the camera's resistance to laser damage. A genetic algorithm is applied to design and optimize a logarithmic phase mask that demonstrates superior protection capability without compromising image quality. Furthermore, to tackle the restoration of complex degraded images after coding, we develop a synthetic WFC dataset and use it to train a U-Net model integrated with the Mamba module. This enhanced architecture combines the efficiency of long-range dependency modeling with the strong local feature extraction ability of Residual blocks. Simulation and experimental results indicate that the wavefront coding system achieves approximately 100 times improvement in protection performance. The proposed decoding algorithm effectively restores image details, outperforming existing methods in imaging quality and significantly improving the overall restoration performance. This study offers a new approach to laser protection for long-wave infrared optical imaging systems operating in laser-rich environments and is expected to have promising applications in the field of infrared imaging system protection.