wf_net: Wavefront post-processing

In mlpeck/zernike: Zernike Polynomials and Fringe Analysis tools

Wavefront post-processing

Description

Perform post-processing of phase and modulation estimates performed by any of the fringe analysis routines

Usage

wf_net(phi, mod, cp, options)

Arguments

phi	wrapped phase map
mod	modulation estimate
cp	a list describing the pupil boundary
options	a list of options. See psfit_options

Details

Called by psifit, vortexfit, fftfit. This now performs all of the post-processing on a calculated wrapped phase and modulus estimate from any of the fringe analysis routines in this package. This includes calculating the actual interferogram location parameters if not specified at call time, cropping if desired, unwrapping the phase, calculating a net and Zernike fit wavefront.

Value

A list with the following components:

phi	the input phase, cropped if requested
mod	the input modulus
cp	interferogram location parameters as returned by circle.pars
cp.orig	input cp
wf.net	Net unsmoothed wavefront; a matrix of class "pupil"
wf.smooth	Net smoothed wavefront
wf.residual	Difference between net wavefront and polynomial fit
fit	Return value from fitzernikes
zcoef.net	Net Zernike coefficients from fit

Note

As of ver. 3.7.2 the values of options$isoseq and cp$obstruct determine what is called to fit Zernikes. If isoseq is TRUE it's either zpm_cart or zapm_cart, with the latter called automatically if cp$obstruct > 0. Otherwise the calls are made to zpm or zapmC.

Author(s)

M.L. Peck [e-mail]

Examples

## illustrates effect of incremental addition of an obstruction
## and standalone use of wf_net
  
require(zernike)
fpath <- file.path(find.package(package="zernike"), "psidata")
files <- scan(file.path(fpath, "files.txt"), what="character")
for (i in 1:length(files)) files[i] <- file.path(fpath, files[i])

# load the images into an array

images <- load.images(files)

# parameters for this run

source(file.path(fpath, "parameters.txt"))

# phase shifts

phases <- wrap((0:(dim(images)[3]-1))/frames.per.cycle*2*pi)
phases <- switch(ps.dir, ccw = -phases, cw = phases, phases)

# target SA coefficients for numerical null.

sa.t <- sconic(diam,roc,lambda=wavelength)
zopt <- psfit_options(satarget=sa.t)
psfit <- psifit(images, phases, psialg="ls", options=zopt)

# get back the raw wavefront

cp <- psfit$cp

eps <- seq(0.1, 0.5, by=0.1)
ne <- length(eps)
pad0 <- rep(0, ne)

## collect some basic data

df.annfits <- data.frame(eps=c(0, eps), sa.null = c(sa.t[1], pad0), sa.obs = c(psfit$fit[9], pad0), sa.net = c(psfit$zcoef.net[8], pad0),
                         rms.net = c(sqrt(crossprod(psfit$zcoef.net)), pad0), defocus=c(psfit$fit[4], pad0))
for (i in seq_along(eps)) {
  cp$obstruct <- eps[i]
  sa.t <- sconic(diam, roc, eps=eps[i], lambda=wavelength)
  zopt$satarget <- sa.t
  fiti <- wf_net(psfit$phi, psfit$mod, cp=cp, options=zopt)
  df.annfits[i+1, 2:ncol(df.annfits)] <- c(sa.t[1], fiti$fit[9], fiti$zcoef.net[8], sqrt(crossprod(fiti$zcoef.net)), fiti$fit[4])
}

rm(eps)
attach(df.annfits)

X11()
plot(eps, sa.null, type="b", ylim=range(df.annfits[,2:ncol(df.annfits)]), xlim=c(0, 0.8), main="Trends with obstruction fraction")
points(eps, sa.obs, type="b", pch=2, col=2, lty=2)
points(eps, sa.net, type="b", pch=3, col=3, lty=3)
points(eps, rms.net, type="b", pch=4, col=4, lty=4)
points(eps, defocus, type="b", pch=5, col=5, lty=5)
grid()
legend(x=0.6, y=0, legend=names(df.annfits)[-1], col=1:5, lty=1:5, pch=1:5)

## plot the first and last smoothed wavefronts to show difference is completely symmetrical

plotn(psfit, fiti, wftype="smooth", qt=c(0,1))
detach(df.annfits)
print(df.annfits, digits=2)

mlpeck/zernike documentation built on Dec. 29, 2024, 7:41 p.m.