Research on Turbulence-Removal Optical Imaging Based on Multi-Scale GAN and Sequential Images

The interference of atmospheric turbulence in the light transmission process can lead to geometric distortions and spatial blurring in images captured by optical imaging systems, severely degrading the quality of optical imaging. This has created technical bottlenecks in critical fields such as military reconnaissance and disaster monitoring, hindering the further development of related technologies. This paper aims to investigate the impact mechanisms of atmospheric turbulence on optical imaging and proposes a turbulence-removal optical imaging method based on Generative Adversarial Networks (GANs) to overcome the limitations imposed by turbulence on optical imaging systems. By introducing a multi-scale convolutional structure into the generator of the GAN, the network model designed in this paper can effectively capture the spatiotemporal characteristics of turbulence-degraded optical images, thereby accurately correcting the geometric distortions and spatial blurring caused by turbulence. The model learns and extracts deep features from turbulence-degraded images to generate high-quality optical imaging results, significantly improving the resolution and visual quality of the imaging system. Experimental results demonstrate that the proposed method exhibits excellent correction performance in both simulated and real turbulence-degraded scenarios, validating its effectiveness and robustness in practical optical imaging applications. This research not only provides a new technical approach for turbulence optical imaging but also offers strong support for addressing the "bottleneck" challenges in related fields, holding significant theoretical and practical value in enhancing the performance of imaging systems.<p></p>