Laser Illuminated Imaging: Multiframe Beam Tilt Tracking and Deconvolution Algorithm.

摘要

A laser-illuminating imaging system operating in the presence of atmospheric turbulence will encounter several sources of noise and diffraction induced errors that can make postprocessing the images difficult. As the beam propagates, turbulence induced tilt will cause the beam to wander off axis from the target. This is especially troublesome in imaging satellites due to most turbulence being closer to the Earth's surface and greatly affecting the beam in the early stages of propagation. Additionally, the returning beam convoluted with the target will encounter turbulence induced tilt that appears as apparent movement of the target between image frames. This results in varying beam intensities at the target surface between imaging frames that can affect registration algorithms. Using simulated data, an algorithm using expectation maximization and least squares techniques was developed that has the ability to estimate both the tilt of the pulsed laser beam and the apparent movement of the object between incoherent frames separately and produce a superior estimate of the target. The results from this algorithm can be used to reduce the effects of beam wander and thus increase the SNR of postprocessed images.