Of all the processing techniques, deconvolution is a real beast. It can work wonders, but it’s also very difficult to master. How many times I just gave up (using some unpolite words) trying to use it?
To my confort, I thought “well, calm down, … my images are noisy and of low quality, and deconvolution doesn’t like that, it only works fine with good images”. But, after some time I got to a conclusion: “If deconvolution ONLY works with good images, … what’s the point? Good images don’t really need deconvolution, do they?”
So, I decided to try again, and to devote time to apply deconvolution to my case. I chose a target which can benefit from deconvolution (the M101 galaxy), and I just took out the dust from the images I got a couple of years ago.
I’m glad to share with you my final result:
And with my conclusions, I wrote this article, hopping it can be helpful to other people out there.
Read the whole article about deconvolution…
M33 is a member of our local group of galaxies. The light of this beast came into my CCD after travelling for about 2,8 milion years.
This is a mosaic of two pieces, as M33 is a wide target. Each one of the pieces consists of LRGB subs of 5 minutes.
Processing has been difficult. As M33 lowered as the imaging session was ongoing, gradients appeared in the subs. Also, the B channel was affected by this lowering. All of this created uneven gradients between frames. As a result, the image appears noisy. Although I’ve tried to fix as much noise as I could, damaging the weakest parts of the galaxy was a potential damage, so I kept noise reduction in somehow a balance.
The core of the galaxy has been cleaned and exposed with wavelets processing. And the color saturation has been pushed to show the active regions in the galaxy. These regions appear as red knots.
Go to this object description and this image technical detail.