ProFound (version 1.14.1)

profoundGetEllipses: Calculate multiple annulus properties of iso-photal ellipses

Description

Returns multiple ellipse properties for an image, assumed to be monotonically decreasing in flux from a bright centre (i.e. a classic galaxy).

Usage

profoundGetEllipses(image = NULL, segim = NULL, segID = 1, levels = 10, magzero = 0,
pixscale = 1, fixcen = TRUE, dobox = FALSE, plot = TRUE, ...)

Arguments

image

Numeric matrix; required, the image we want to analyse.

segim

Integer matrix; optional, the segmentation map of the image. This matrix *must* be the same dimensions as image.

segID

Integer scalar; optional, the desired segim segment to extract from the image.

levels

Integer scalar or vector. If a scalar this is the number of ellipse levels to extract from the image. If a vector this specfies the extremes of all fractional levels, i.e. it should generally start at 0 and end at 1 to capture all isophotal levels.

magzero

Numeric scalar; the magnitude zero point. What this implies depends on the magnitude system being used (e.g. AB or Vega). If provided along with pixscale then the surface brightness output will represent mag/asec^2.

pixscale

Numeric scalar; the pixel scale, where pixscale=asec/pix (e.g. 0.4 for SDSS). If set to 1 (default), then the output radhi, radlo and radav is in terms of pixels, otherwise they are in arcseconds. If provided along with magzero then the surface brightness output will represent mag/asec^2.

fixcen

Logical; should the ellipse centres be fixed to a common flux weighted centre?

dobox

Logical; should boxiness be computed? If FALSE then boxiness is fixed to be 0. If TRUE then boxiness is computed (and other parameters are refined) using a maximum likelihood method. This is more expensive to compute, so the default is FALSE.

plot

Logical; should a diagnostic plot be generated? This plot is generated by the profoundGetEllipsesPlot function.

…

Further arguments to be passed to profoundGetEllipsesPlot. Only relevant if plot=TRUE.

Value

A list containing:

ellipses

A data.frame of ellipse properties ordered by radius (see below).

segellipses

Integer matrix; the ellipse-wise segmentation map matched pixel by pixel to image. This allows you to see which specific pixels used to compute each ellipse annulus in ellipses, where the number in the segmentation map refers to segellipseID.

ellipses is a data.frame of ellipse properties ordered by radius. It has the following columns

segellipseID

The ellipse segment ID that refers to the segmentation map segellipses.

fluxfrac

The approximate fraction of galaxy flux contained within this ellipse.

xcen

The flux weighted x centre of the ellipse.

ycen

The flux weighted y centre of the ellipse.

radhi

The major axis extent of the ellipse (units of pixscale, so if pixscale represents the standard asec/pix this will be asec)..

radlo

The minor axis extent of the ellipse (units of pixscale, so if pixscale represents the standard asec/pix this will be asec)..

radav

The average radius of the ellipse (units of pixscale, so if pixscale represents the standard asec/pix this will be asec)..

axrat

The axial ratio of the ellipse as given by radlo/radhi.

ang

The angle of the ellipse in the usual ProFit sense, see profitMakeModel.

box

The boxiness of the ellipse in the usual ProFit sense, see profitMakeModel.

xsd

The flux weighted standard deviation in x (always in units of pix).

ysd

The flux weighted standard deviation in y (always in units of pix).

covxy

The flux weighted covariance in xy (always in units of pix).

corxy

The flux weighted correlation in xy (always in units of pix).

flux

The flux contained in the segmented pixels associated with this ellipse.

N

The number of segmented pixels associated with this ellipse.

SB

The mean surface brightness of the pixels associated with this ellipse (if pixscale has been set correctly then this column will represent mag/asec^2, otherwise it will be mag/pix^2).

Details

This higher level function provides an easy way to extract iso-photal ellipses from an image of a galaxy. How it works somewhat replicates IRAF's ellipse, but it is really present to offer useful initial guesses for bulge and disk geometric properties. It certainly does not guarantee to return the same solution as IRAF (in fact I am not exactly aware of how IRAF computes its ellipses).

Internally it works by rank ordering the pixels of the galaxy and dividing these into equi-spaced quantiles of flux (so each annulus will approximately sum to the same amount of flux). This means that the error for each ellipse will be approximately constant. For each annulus it then runs profoundGetEllipse to compute the ellipse properties of what is assumed to be a fairly narrow annulus of pixels. The implicit assumption is that the galaxy flux more-or-less monotonically decreases from the centre, and dividing pixels like this will assure the extraction of common iso-photal ellipses. This assumption works well within the inner 90% of a galaxy's flux, but isophotes can be quite noisy once the galaxy flux gets close to the sky RMS level. This said, the ellipse returned will on average make sense, and ellipses tend to overlap only in very extreme cases (where the geometry is highly non-elliptical or there are close contaminants).

See Also

profoundGetEllipsesPlot, profoundGetEllipse, profoundDrawEllipse

Examples

# NOT RUN {
# We need the ProFit library to show the profile: library(ProFit)
image = readFITS(system.file("extdata", 'KiDS/G278109fitim.fits',
package="ProFit"))$imDat
segim = readFITS(system.file("extdata", 'KiDS/G278109segim.fits',
package="ProFit"))$imDat
ellipses_nobox = profoundGetEllipses(image=image, segim=segim, levels=20, dobox=FALSE,
pixscale=0.2)
ellipses_box = profoundGetEllipses(image=image, segim=segim, levels=20, dobox=TRUE,
pixscale=0.2)

magplot(ellipses_box$ellipses$radhi[4:19], ellipses_nobox$ellipses$SB[4:19],
ylim=c(25,17), grid=TRUE, type='l')
points(ellipses_box$ellipses$radhi[4:19],ellipses_box$ellipses$SB[4:19])
#A rough bulge+disk surface brightness profile (mean axrat~0.6):
rlocs=seq(1,30,by=0.1)
bulge=profitRadialSersic(rlocs, mag=18.2, re=1.7, nser=3)
disk=profitRadialSersic(rlocs, mag=18, re=13, nser=0.7)
lines(rlocs, profoundFlux2SB(bulge, pixscale=0.2), col='red')
lines(rlocs, profoundFlux2SB(disk, pixscale=0.2), col='blue')
lines(rlocs, profoundFlux2SB(bulge+disk, pixscale=0.2), col='green')
#To get correct magnitudes you would need to modify the components by the axrat
#and pixel scale.

#We can do a better 1D fit with ease:
#Since the ellipses are divided by equi-flux we can minimise sum-square of the SB diff:
sumsq1D=function(par=c(17.6, log10(1.7), log10(3), 17.4, log10(13), log10(0.7)),
rad, SB, pixscale=1){
  bulge=profitRadialSersic(rad, mag=par[1], re=10^par[2], nser=10^par[3])
  disk=profitRadialSersic(rad, mag=par[4], re=10^par[5], nser=10^par[6])
  total=profoundFlux2SB(bulge+disk, pixscale=pixscale)
  return=sum((total-SB)^2)
}

lower=c(10,0,-0.5,10,0,-0.5)
upper=c(30,2,1,30,2,1)

fit1D=optim(sumsq1D, par=c(17.6, log10(1.7), log10(3), 17.4, log10(13), log10(0.7)),
rad=ellipses_box$ellipses$radhi[4:19], SB=ellipses_box$ellipses$SB[4:19], pixscale=0.2,
method='L-BFGS-B', lower=lower, upper=upper)$par

magplot(ellipses_box$ellipses$radhi[4:19], ellipses_nobox$ellipses$SB[4:19],
ylim=c(25,17), grid=TRUE, type='l')
points(ellipses_box$ellipses$radhi[4:19],ellipses_box$ellipses$SB[4:19])
#A simple bulge+disk surface brightness profile:
rlocs=seq(1,30,by=0.1)
bulge=profitRadialSersic(rlocs, mag=fit1D[1], re=10^fit1D[2], nser=10^fit1D[3])
disk=profitRadialSersic(rlocs, mag=fit1D[4], re=10^fit1D[5], nser=10^fit1D[6])
lines(rlocs, profoundFlux2SB(bulge, pixscale=0.2), col='red')
lines(rlocs, profoundFlux2SB(disk, pixscale=0.2), col='blue')
lines(rlocs, profoundFlux2SB(bulge+disk, pixscale=0.2), col='green')
# }