Horsehead Nebula (Barnard 33)

This astrophotograph shows the Horsehead Nebula (catalogued as Barnard 33), and the Flame Nebula (catalogued as NGC 2024). The Horsehead Nebula is a dark nebula silhouetted against background nebulosity. The Flame Nebula is an emission nebula. They are situated next to the easternmost star in the Belt of Orion, which is the brightest star in this photo.

Light pollution

This photograph was taken in lots of light pollution: There was a nearly-full moon (if moonlight can be called light pollution!), and in addition, my neighbours had their wall-mounted stadium light on full blast. It was shining almost directly down my telescope. I had to take care to correct the nasty gradient resulting from this light pollution. The Altair Quad Band Filter really helped, and I’m quite pleased with this result.

Lots of light pollution

New kit

After exactly 3 months of cloud (including just one other clear night that I couldn’t go out for), this was long-awaited! I also had two new items of kit to try out, acquired during the 3-month ‘astro drought’:

  1. I bought this filter tray and screwed it directly into the camera’s T2 thread, putting the filter much closer to the sensor than before. I then reconfigured the spacers between the flattener and the sensor to maintain exactly the required 115 mm. This ensures nice round stars, and putting the filter right next to the sensor should help reduce halos and reflections.
  2. I also bought this cheap light box off Amazon to save me having to take flat frames against the dawn sky. When combined with two folds of a white T-shirt to disperse the light, it works very well. I simply stretched the folded T-shirt over the dew shield, turned the telescope vertical, and rested the light box on top.
Taking flat frames with a light box

Converting bayer patterns

I did run into one unexpected problem: The lightbox is very bright, and so I only needed 0.5-s exposures. For exposures this short, SharpCap automatically switches from FITS format to SER, and my stacking software, DeepSkyStacker, does not read SER files. I thought, “no problem, I’ll use PIPP to convert them back to FITS”. DeepSkyStacker, however, reads FITS in top-bottom format, and PIPP reads them in bottom-top format. That’s the difference between an RGGB and a BGGR bayer mosaic. My flat frames would not work and instead damaged the image. Amazingly, I could find no software anywhere to flip the flat frames vertically and maintain the RGGB bayer mosaic, keeping them consistent with my light frames.

Rather than simply retake the flat frames in the correct mode, I instead used Python and the astropy library to fix my flat frames. Problem solved. Next time I will try to remember to check I have the FITS file format selected.

My code is functional and not particularly pretty (looping, eugh). Please don’t judge me. Nevertheless, here it is in case it can be useful to anyone:

import os
from import fits
import cv2 # pip install opencv-python
import glob

if __name__ == "__main__":

	fits_files = glob.glob('*.fit')

	for fits_file in fits_files:
		hdu =[0]
		image_data =
		for i in range(0,hdu.header['NAXIS2']-1,2):
			for j in range(0,hdu.header['NAXIS1']-1,2):
		image_data = cv2.flip(image_data, 0)
		hdu = fits.PrimaryHDU(image_data)
		hdu.writeto(os.path.splitext(fits_file)[0] +'-rebayered.fits',overwrite=True)


  • 29× 300-s light frames (Gain 900)
  • Full use of calibration frames (darks, flats and dark flats)


  • Explore Scientific ED 102 mm Apo f/7 refractor
  • Sky-Watcher EQ6-R PRO SynScan GOTO equatorial mount
  • Altair Hypercam 294C PRO colour fan-cooled camera
  • Altair quad-band one-shot colour (OSC) 2″ filter
  • Revelation Adjustable Field Flattener
  • Altair 60mm guide scope
  • Altair GPCAM2 AR0130 mono guide camera


  • Sharpcap
  • PHD2
  • DeepSkyStacker
  • Photoshop
  • StarNet
  • Topaz Labs DeNoise AI

Leave a comment