ProFound (version 1.14.1)

FPtest: False Positive Reference Data

Description

This data consists of 1,000 runs of a random 1000 x 1000 noise matrix through profoundProFound. The catalogue is a concatenation of all the segstats outputs for all of these run.

Usage

data("FPtest")

Format

A data frame with 7012 observations on the following 56 variables. See profoundProFound for a detailed discussion on each of these parameters.

segID

a numeric vector

uniqueID

a numeric vector

xcen

a numeric vector

ycen

a numeric vector

xmax

a numeric vector

ymax

a numeric vector

RAcen

a logical vector

Deccen

a logical vector

RAmax

a logical vector

Decmax

a logical vector

sep

a numeric vector

flux

a numeric vector

mag

a numeric vector

cenfrac

a numeric vector

N50

a numeric vector

N90

a numeric vector

N100

a numeric vector

R50

a numeric vector

R90

a numeric vector

R100

a numeric vector

SB_N50

a numeric vector

SB_N90

a numeric vector

SB_N100

a numeric vector

xsd

a numeric vector

ysd

a numeric vector

covxy

a numeric vector

corxy

a numeric vector

con

a numeric vector

asymm

a logical vector

flux_reflect

a logical vector

mag_reflect

a logical vector

semimaj

a numeric vector

semimin

a numeric vector

axrat

a numeric vector

ang

a numeric vector

signif

a numeric vector

FPlim

a numeric vector

flux_err

a numeric vector

mag_err

a numeric vector

flux_err_sky

a numeric vector

flux_err_skyRMS

a numeric vector

flux_err_shot

a numeric vector

sky_mean

a numeric vector

sky_sum

a numeric vector

skyRMS_mean

a numeric vector

Nedge

a logical vector

Nsky

a logical vector

Nobject

a logical vector

Nborder

a logical vector

Nmask

a logical vector

edge_frac

a logical vector

edge_excess

a logical vector

flag_border

a logical vector

iter

a numeric vector

origfrac

a numeric vector

flag_keep

a logical vector

Details

Specifically we ran with defaults the following command 1,000 times in a loop:

profoundProFound(matrix(rnorm(1e6),1e3))

The output is then a reference of the false positive rate, since we have not injected any sources into the images. The fact we find 7,012 false detections mean we expect 7 false positives per 1e6 pixels (the size in pixels of the input matrix). To compare against any target data we need to adjust the magnitudes by the sky RMS magnitude level, i.e. add on profoundFlux2Mag(skyRMS, 0) (if the zero point is 0 for our target data). See Examples for a comparison to our included VIKING data.

Examples

# NOT RUN {
image=readFITS(system.file("extdata", 'VIKING/mystery_VIKING_Z.fits', package="ProFound"))
profound=profoundProFound(image, magzero=30, rotstats=TRUE)
skyRMS=median(profound$skyRMS)
magoff=profoundFlux2Mag(skyRMS, 30)
totpix=prod(profound$dim)

#We can easily compute the expected number of false positives on an image this size:
data("FPtest")
dim(FPtest)[1]*totpix/1e6/1e3

#And plot the detections and expected false positive distributions:
maghist(profound$segstats$mag, seq(-11,-1,by=0.2)+magoff)
maghist(FPtest$mag+magoff, seq(-6,-1,by=0.2)+magoff, scale=totpix/1e6/1e3, add=TRUE,
border='red')
# }