Proceedings Article | 21 August 2013
KEYWORDS: Reconstruction algorithms, Computer simulations, Optical simulations, Atmospheric propagation, Laser beam propagation, Fourier transforms, Interference (communication), Turbulence, Radar, Space reconnaissance
Currently simulating on laser propagation in atmosphere usually need to deal with beam in strong turbulence,
which may lose a part of information via Fourier Transform to simulate the transmission, makes the phase of beam as a
2-D array wrap by 2π . An effective unwrapping algorithm is needed for continuing result and faster calculation. The
unwrapping algorithms in atmospheric propagation are similar to the unwrapping algorithm in radar or 3-D surface
rebuilding, but not the same. In this article, three classic unwrapping algorithms: the block least squares (BLS), mask-cut
(MCUT), and the Flynn’s minimal discontinuity algorithm (FMD) are tried in wave-front reconstruction simulation.
Each of those algorithms are tested 100 times in 6 same conditions, including low(64x64), medium(128x128), and
high(256x256) resolution phase array, with and without noises. Compared the results, the conclusions are delivered as
follows. The BLS-based algorithm is the fastest, and the result is acceptable in low resolution environment without noise.
The MCUT are higher in accuracy, though they are slower with the array resolution increased, and it is sensitive to noise,
resulted in large area errors. Flynn’s algorithm has the better accuracy, and it occupies large memory in calculation. After
all, the article delivered a new algorithm that based on Active on Vertex (AOV) Network, to build a logical graph to cut
the search space then find minimal discontinuity solution. The AOV is faster than MCUT in dealing with high resolution
phase arrays, and better accuracy as FMD that has been tested.
