Beyond the current noise limit in imaging through turbulent medium

Area of Science:

• Optics and Photonics
• Image Processing
• Computational Imaging

Background:

• Image resolution degrades due to atmospheric turbulence.
• Traditional shift-and-add methods struggle with low signal-to-noise ratios.
• Object detection is challenging when fainter than background noise.

Purpose of the Study:

• To develop a highly efficient method for determining frame shifts in shift-and-add imaging.
• To improve image reconstruction from data with low object-to-background contrast.
• To extend the applicability of shift-and-add techniques to challenging imaging scenarios.

Main Methods:

• A custom genetic algorithm was developed for iterative shift estimation.
• Maximal energy of stacked images was used as the objective function.
• The method was validated on simulated and real images of varying complexity.

Main Results:

• The genetic algorithm successfully determined frame shifts even when objects were only 2% brighter than background noise.
• High-resolution, high signal-to-noise ratio images were reconstructed.
• The method demonstrated effectiveness on both simple and complex sources.

Conclusions:

• The proposed method significantly advances image reconstruction limits using shift-and-add.
• It enables recovery of spatial distributions in extremely low light conditions.
• Applications span optical and infrared imaging, including digital image stabilization.