Notebook

`astropy.io.fits`¶

In astropy, one of the most important function is fits, which is used for reading and writing fits file. ccdproc or any other packages which loads fits image is strongly recommended to use astropy.io.fits. This section is a summary of how to use this function.

As of Astropy 2.0, the CCDData class of ccdproc has implemented to astropy.nddata. So the previous

from ccdproc import CCDData

is identical to

from astropy.nddata import CCDData

The ccdproc.CCDData may be deprecated soon by issue #510.

In [3]:

from astropy.io import fits

hdul = fits.open('HST_Tutorial/Pluto1.fits')
# hdul = HDU List = Header Data Unit List

print(type(hdul))
print()
hdul.info()
# prints out the information of the hdul.

<class 'astropy.io.fits.hdu.hdulist.HDUList'>

Filename: HST_Tutorial/Pluto.fits
No.    Name         Type      Cards   Dimensions   Format
  0  PRIMARY     PrimaryHDU     314   ()      
  1  SCI         ImageHDU       175   (513, 512)   float32   
  2  ERR         ImageHDU        51   (513, 512)   float32   
  3  DQ          ImageHDU        43   (513, 512)   int16   
  4  D2IMARR     ImageHDU        15   (64, 32)   float32   
  5  D2IMARR     ImageHDU        15   (64, 32)   float32   
  6  WCSDVARR    ImageHDU        15   (64, 32)   float32   
  7  WCSDVARR    ImageHDU        15   (64, 32)   float32   
  8  WCSCORR     BinTableHDU     59   7R x 24C   [40A, I, A, 24A, 24A, 24A, 24A, D, D, D, D, D, D, D, D, 24A, 24A, D, D, D, D, J, 40A, 128A]

TIP: You can use fitsinfo, fitsheader, etc on terminal, too!

You may wonder what the PRIMARY, SCI, ect, are. These depend on the instrument you used for the observation; two most important things are PRIMARY and SCI, which means the primary (common) header and scientific image, respectively. I intentionally downloaded raw data, i.e., no much correction is done to the image (e.g., cosmic ray rejection).

Basically there are two important information in HDUList: header and data (as its name indicates). You can check both by :

In [8]:

# Let me use SCI (index 1) only
header = hdul[1].header 
data   = hdul[1].data
print('which ccd chip is this? ', header['ccdchip'])

which ccd chip is this?  2

You can check the whole header by print(header). What about data? In the following, I will show you how to plot the "zscale" image:

In [12]:

from matplotlib import pyplot as plt
from astropy.visualization import ZScaleInterval, ImageNormalize

def znorm(image, **kwargs):
    return ImageNormalize(image, interval=ZScaleInterval(**kwargs))

def zimshow(image, **kwargs):
    plt.imshow(image, norm=znorm(image, **kwargs), origin='lower')
    plt.colorbar()
    
zimshow(data)
plt.show()

Or you can just use ginga or ds9.

Crop the Image¶

You can crop the image using various ways. One possible, and probably the most useful way is to use ccdproc. But this ccdproc is using some variant of fits: CCDData. This is used for pythonic image processing. So I will load the image in CCDData, not HDUList format:

In [15]:

from ccdproc import CCDData, trim_image
import numpy as np

data = CCDData(data = hdul[1].data, 
               header=hdul[0].header+hdul[1].header, 
               unit='adu')

# FITS-style:
data1 = trim_image(data, fits_section='[1:200, 1:100]')

# python-style:
data2 = trim_image(data[:100, :200])
# 0:100 means 0 to 99 in Python, so that the total number is same as fits (100 pixels)

print(np.sum(data1.subtract(data2)) == 0)

# data2.write('test.fits', overwrite=True)

True

Why did I use header = hdul[0].header + hdul[1].header ? That is because hdul[0] contains the common header information, so if I skip this part, the header from hdul[1] is not perfect. If you try to use only hdul[1].header, you may get an error, or the output file (test.fits) may not contain any wcs information.

Also by using fits_section, you can freely choose Python convention or FITS convention for trimming the image.

astropy.io.fits¶

Crop the Image¶

`astropy.io.fits`¶