Turbulence video simulation using the spatiotemporal cross correlation of distortion fields

Area of Science:

• Computer Vision
• Optical Engineering
• Fluid Dynamics

Background:

• Simulating atmospheric turbulence in videos is computationally intensive.
• Existing methods often focus on single images, lacking temporal dynamics.

Purpose of the Study:

• Develop an efficient algorithm for simulating spatiotemporal videos affected by atmospheric turbulence.
• Incorporate temporal turbulence properties and blurring effects into existing single-image simulation methods.

Main Methods:

• Extend single-image turbulence simulation by adding time-domain properties and blurring.
• Analyze spatiotemporal correlations of turbulence-induced image distortions.
• Incorporate turbulence strength, object distance, and height as input parameters.

Main Results:

• Successfully simulated spatiotemporal videos with atmospheric turbulence.
• Validated that simulated video distortion fields match physical spatiotemporal cross-correlation functions.
• Demonstrated applicability to both low and high frame rate videos.

Conclusions:

• The developed algorithm provides an efficient method for simulating turbulence in videos.
• This simulation is valuable for training algorithms that process turbulence-degraded imaging data.
• The ease of use and accurate correlation modeling make it a practical tool for research and development.