view article

Figure 2
(a) Block diagram of the modified Gerchberg–Saxton algorithm with modulus and bandwidth constraint. O denotes the object and D the detector space. (b) Reconstructed relative phase ϕ and beam profile (normalized amplitude) along the sample from simulated data. The starting guess (blue, dotted) was taken to be constant but can also be a random guess. (c) Forward-simulated diffraction pattern, corresponding to the input data of the simulation, as calculated by a Fresnel–Kirchhoff integral, compared with the diffraction data of the final reconstruction. Even though the beam profile (illumination function) B(x) is unconstrained (except by reality), it is well reconstructed apart from minor distortions. The SAW pattern is well reconstructed in view of its magnitude, and the reconstructed profile is in exactly the right position with respect to the beam, i.e. its envelope function. (d) Logarithmic error metric log10(E) shown for two simulated data sets (noise-free and Poisson noise, expectation value [\gamma=3.21\times 10^{5}]).

Journal logoJOURNAL OF
APPLIED
CRYSTALLOGRAPHY
ISSN: 1600-5767
Follow J. Appl. Cryst.
Sign up for e-alerts
Follow J. Appl. Cryst. on Twitter
Follow us on facebook
Sign up for RSS feeds