interp.2d: Interpolate 2D function

In AngusWright/LAMBDAR: Measure galaxy fluxes from an arbitrary FITS image using input aperture locations and dimensions

Interpolate 2D function

Description

An interpolation function for 2D distributions

Usage

interp.2d(x,y,obj)

Arguments

x	numeric (n) array; x-indices of destination grid
y	numeric (n) array; y-indices of destination grid
obj	list; elements x: the x indices of the original grid; y: the y indices of the original grid; zobj: the original 2D image matrix

Details

Function is used for interpolating images from one grid onto a new grid of arbitrary resolution. The function is not able to perform rotations etc, but is particularly useful for modifying astronomical images to match each other in pixel space.

Value

length(n,3) data.frame; x indicies, y indicies, and the interpolated image values at (x,y).

Author(s)

Angus H Wright ICRAR [e-mail];

Aaron Robotham ICRAR [e-mail]

Examples

#Load LAMBDAR
library(LAMBDAR)

#Generate the original grid/image
x.min<-y.min<--50
x.max<-y.max<-50
x<-seq(x.min,x.max)
y<-seq(y.min,y.max)
old.grid<-expand.grid(x,y)
#Make image matrix of r values
z<-matrix(sqrt(old.grid[,1]^2+old.grid[,2]^2),length(x),length(y))

#Generate the new grid to interpolate onto
x.new.min<-0
y.new.min<--15
x.new.max<-25
y.new.max<-15
x.new<-seq(x.new.min,x.new.max,by=0.1)
y.new<-seq(y.new.min,y.new.max,by=0.1)
new.grid<-expand.grid(x.new,y.new)

#Interpolate
z.new<-interp.2d(x=new.grid[,1],y=new.grid[,2],obj=list(x=x,y=y,z=z))

#Reconstruct image
z.new<-matrix(z.new[,3],length(x.new),length(y.new))

#Plot the two images
layout(cbind(1,2))
image(x=x,y=y,z=z,col=rainbow(1E1),xlab='X (pix)',
ylab='Y (pix)',xlim=c(x.min,x.max)*2/3,ylim=c(x.min,x.max)*2/3,
asp=1,zlim=c(0,max(z)/2))
image(x=x.new,y=y.new,z=z.new,col=rainbow(1E1),zlim=c(0,max(z)/2),
add=TRUE)
rect(x.new.min,y.new.min,x.new.max,y.new.max,col='gold',density=0,
lwd=3)

#Using this implementation, the function can also be used
#to interpolate and sample from a function simultaneously.

#Generate sampling points (in the middle 2/3 of the image)
x.samp<-runif(350,min=x.min,max=x.max)*2/3
y.samp<-runif(350,min=y.min,max=y.max)*2/3
samp.grid<-cbind(x.samp,y.samp)

#Interpolate
z.samp<-interp.2d(x=samp.grid[,1],y=samp.grid[,2],obj=list(x=x,y=y,z=z))

#Reconstruct image
z.samp.im<-matrix(NA,length(x),length(y))
z.samp.im[cbind(x.samp-x.min+1,y.samp-y.min+1)]<-z.samp[,3]

#Plot the image
image(x=x,y=y,z=z,col=grey.colors(1E1),xlab='X (pix)',
ylab='Y (pix)',xlim=c(-50,50),ylim=c(-50,50),zlim=c(0,max(z)/2),asp=1)
image(x=x,y=y,z=z.samp.im,col=rainbow(1E1),zlim=c(0,max(z)/2),add=TRUE)

AngusWright/LAMBDAR documentation built on Sept. 19, 2024, 10:26 a.m.