Air Turbulence Mitigation Techniques for Long-Range Terrestrial Surveillance

Abstract

Atmospheric turbulence severely reduces the effective range of the visual surveillance due to geometric deformation and contrast degradation. This paper for the first time puts together various methodologies used for mitigating the air turbulence and the challenges associated with them. Historically, atmospheric turbulence was studied for astronomical observations and remote sensing; quite recently, it is revisited for long-range terrestrial (horizontal) surveillance (LRTS). LRTS has a diverse range of applications such as border security, long-distance photography, identification, detection of distant objects, etc. Atmospheric turbulence mitigation is a non-trivial problem due to spatial and temporal fluctuations in the air refractive index. The existing approaches for air turbulence mitigation are lucky imaging, adaptive optics, averaging, deblurring, and image registration-based methods. In long-range imaging systems, image registration-based turbulence mitigation techniques are the most popular, but these methods distort the independent motion existing in the scene. In the recent past, few approaches have been introduced to reduce the effect of air turbulence while preserving the object motion. Usually video stabilization techniques yield blurring of output frames, so deblurring and post-processing are necessary steps for better visualization.