Comparison between the plenoptic sensor and the light field camera in restoring images through turbulence

Similar to the "lucky imaging" technique that selects the best local features over time, spatial redundancy allows for the localization of turbulence induced image distortions, and selection of the best features which are least distorted by turbulence effect. A new technique to restore turbulence degraded images is proposed based on imaging with spatial redundancies. Two imaging frameworks which are candidates for implementation of the technique are the plenoptic sensor and the light field camera, which collect multiple depictions of the target through sub-aperture imaging. Preliminary studies have demonstrated the effectiveness of either device in imaging through turbulence. However, as visual distortions vary significantly from weak to strong turbulence conditions, it is unclear when and how a light field approach should be applied to enhance target recognition over distorted medias. We present an in-depth study on the fundamental differences between the two devices with regards to turbulence distortion, as well as their image restoration mechanisms. Our analysis combined with proof-of-concept experiments show that the turbulence resilience of light field imaging techniques depends strongly on the mechanism of mapping the light field. Such universal finding serves as guidance for imaging and object recognition with light field approaches.