Long-distance SWIR turbulence mitigation algorithm based on fast registration and multilevel restoration

Short-wave infrared (SWIR) imaging systems play a pivotal role in military, industrial, and environmental monitoring applications due to their exceptional atmospheric penetration capability and all-weather detection performance. This study addresses the critical challenge of turbulence-induced degradation in long-range SWIR imaging under complex atmospheric conditions by proposing a novel turbulence suppression algorithm integrating fast registration and multi-stage restoration. The proposed framework comprises four key technical components: (1) a frequency-aware reference frame selection mechanism, (2) an iterative hexagonal block matching strategy for rapid inter-frame alignment, (3) an adaptive deblurring module combining Wiener filtering with lucky region fusion, and (4) a Gaussian difference of Gaussian (DoG) operator for detail enhancement. Experimental results demonstrate that our algorithm effectively suppresses turbulenceinduced geometric distortions while significantly enhancing image details, achieving remarkable improvements in both subjective visual quality and objective metrics. This work provides a new technical solution for infrared surveillance in turbulent atmospheric environments.