Subsetting
First we will load up an image in RAM (temp_image) and as a pointer
to the disk object (temp_point)
file_image = system.file('extdata', 'image.fits', package = "Rfits")
temp_image = Rfits_read_image(file_image)
temp_point = Rfits_point(file_image)
We can look at both of these, since they have distinct print methods
by default:
print(temp_image)
#> Class: Rfits_image
#> File path: /Library/Frameworks/R.framework/Versions/4.2/Resources/library/Rfits/extdata/image.fits
#> Ext num: 1
#> Ext name:
#> RAM size: 0.9871 MB
#> BITPIX: -32
#> NAXIS1: 356
#> NAXIS2: 356
#> Key N: 23
print(temp_point)
#> File path: /Library/Frameworks/R.framework/Versions/4.2/Resources/library/Rfits/extdata/image.fits
#> Ext num: 1
#> Ext name:
#> Class: Rfits_pointer
#> Type: image
#> Dim: 356 356
#> Disk size: 0.4889 MB
#> BITPIX: -32
#> Key N: 23
The two data structures are a bit different, but we can still plot
both directly:
plot(temp_image)
plot(temp_point)
Behind the scene the S3 plot method does different things depending
on the Rfits object encountered. If it is an Rfits_image then it will
directly plot it using Rwcs_image with a WCS. If it is a
Rfits_pointer then it will first fully load the data into RAM, and then
plot it using Rwcs_image.
If you have a Rfits_pointer object then you can convert it to a
Rfits_image very easily:
temp_point[]
#> Class: Rfits_image
#> File path: /Library/Frameworks/R.framework/Versions/4.2/Resources/library/Rfits/extdata/image.fits
#> Ext num: 1
#> Ext name:
#> RAM size: 0.9871 MB
#> BITPIX: -32
#> NAXIS1: 356
#> NAXIS2: 356
#> Key N: 23
This is because Rfits has special subset operation that work
specifically on Rfits_image and Rfits_pointer objects. This means we can
plot a subset:
plot(temp_image[50:150,50:150])
plot(temp_point[50:150,50:150])
For convenience the following will achieve the same outcomes:
plot(temp_image[c(50,150),c(50,150)])
plot(temp_point[c(50,150),c(50,150)])
And we can also cutout a box on a specific location:
plot(temp_image[101, 101, box=100])
plot(temp_point[101, 101, box=100])
We can also subset based on RA/Dec coordinates:
coords = Rwcs_p2s(101,101,temp_image$keyvalues)
print(coords)
#> RA Dec
#> [1,] 352.3 -31.8296
plot(temp_image[coords[1], coords[2], box=100, type='coord'])
plot(temp_point[coords[1], coords[2], box=100, type='coord'])
Useful Class Methods
Rfits comes with a whole lot of specific Rfits_image and
Rfits_pointer class methods (for details see ?Rfits_methods):
- print
- length (total length of the image/cube/array)
- dim (dimension of the image/cube/array)
- centre/center (RA, Dec of image centre)
- corners (RA, Dec of image corners)
- extremes (RA, Dec of image extremes)
- pixscale (pixel scale of centre of image)
- pixarea (pixel area of centre of image)
- rotation (rotation of image in degrees)
They are designed to give identical results, i.e.:
print(temp_image)
#> Class: Rfits_image
#> File path: /Library/Frameworks/R.framework/Versions/4.2/Resources/library/Rfits/extdata/image.fits
#> Ext num: 1
#> Ext name:
#> RAM size: 0.9871 MB
#> BITPIX: -32
#> NAXIS1: 356
#> NAXIS2: 356
#> Key N: 23
print(temp_point)
#> File path: /Library/Frameworks/R.framework/Versions/4.2/Resources/library/Rfits/extdata/image.fits
#> Ext num: 1
#> Ext name:
#> Class: Rfits_pointer
#> Type: image
#> Dim: 356 356
#> Disk size: 0.4889 MB
#> BITPIX: -32
#> Key N: 23
length(temp_image)
#> [1] 126736
length(temp_point)
#> [1] 126736
dim(temp_image)
#> [1] 356 356
dim(temp_point)
#> [1] 356 356
centre(temp_image)
#> RA Dec
#> [1,] 352.2914 -31.8223
centre(temp_point)
#> RA Dec
#> [1,] 352.2914 -31.8223
corners(temp_image)
#> RA Dec
#> BL 352.3111 -31.83906
#> TL 352.3111 -31.80554
#> TR 352.2717 -31.80554
#> BR 352.2717 -31.83906
corners(temp_point)
#> RA Dec
#> BL 352.3111 -31.83906
#> TL 352.3111 -31.80554
#> TR 352.2717 -31.80554
#> BR 352.2717 -31.83906
extremes(temp_image)
#> RA Dec
#> min 352.2717 -31.839059
#> max 352.3111 -31.805535
#> range 2.0114 2.011401
extremes(temp_point)
#> RA Dec
#> min 352.2717 -31.839059
#> max 352.3111 -31.805535
#> range 2.0114 2.011401
pixscale(temp_image)
#> [1] 0.3390001
pixscale(temp_point)
#> [1] 0.3390001
pixarea(temp_image)
#> [1] 0.114921
pixarea(temp_point)
#> [1] 0.114921
rotation(temp_image)
#> North East
#> [1,] 0 180
rotation(temp_point)
#> North East
#> [1,] 0 180
Pixel Manipulation Methods
Rfits_image objects can be usefully operated on as if they were
matrices:
plot(temp_image^5, qdiff=TRUE)
plot(temp_image - 10, qdiff=TRUE)
Most R mathematical operators will work. You can
also directly operate with Rfits_image and Rfits_point objects:
plot(temp_image * temp_image, qdiff=TRUE)
plot(temp_image - temp_image, qdiff=TRUE) #blank, since the image should now be 0
plot(temp_point - temp_point, qdiff=TRUE) #blank, since the image should now be 0
If you mix Rfits object types you will need to convert the
Rfits_pointer to a Rfits_image using the subset operation:
plot(temp_image * temp_point[], qdiff=TRUE)
plot(temp_image - temp_point[], qdiff=TRUE) #blank, since the image should now be 0