Single Frame Atmospheric Turbulence Mitigation: A Benchmark Study and A New Physics-Inspired Transformer Model
This work addresses the problem of image restoration for applications like surveillance or astronomy by introducing a novel transformer-based approach and providing new datasets, though it is incremental in building upon existing deep learning methods.
The paper tackles the challenge of restoring images distorted by atmospheric turbulence, which involves spatially varying blur and geometric distortions, by proposing a physics-inspired transformer model that jointly estimates a dynamic distortion map and recovers a turbulence-free image, achieving improved performance on new real-world datasets with metrics like PSNR and SSIM.
Image restoration algorithms for atmospheric turbulence are known to be much more challenging to design than traditional ones such as blur or noise because the distortion caused by the turbulence is an entanglement of spatially varying blur, geometric distortion, and sensor noise. Existing CNN-based restoration methods built upon convolutional kernels with static weights are insufficient to handle the spatially dynamical atmospheric turbulence effect. To address this problem, in this paper, we propose a physics-inspired transformer model for imaging through atmospheric turbulence. The proposed network utilizes the power of transformer blocks to jointly extract a dynamical turbulence distortion map and restore a turbulence-free image. In addition, recognizing the lack of a comprehensive dataset, we collect and present two new real-world turbulence datasets that allow for evaluation with both classical objective metrics (e.g., PSNR and SSIM) and a new task-driven metric using text recognition accuracy. Both real testing sets and all related code will be made publicly available.