SharpCap Forums

Captured 9th July 2022 @ 21:30.



Autostakkert output - best 10% of 1700 frames.



The red areas are the inverted mask which protects brighter parts of the image from further processing.



For PSF, Standard Deviation and Iterations were arrived at by trial and error.




Dave

I have added named craters - info from Virtual Moon Atlas.



Dave

Really good Dave, Exposure and detail looks great across the entire moon .

So this is a move away from using AS! and Registax to PI for processing lunar images? I was trying to follow what you have done since have yet to try PI for moon and planets - a bit different from deep sky.

The red filter presumably improves contrast for the bright moon?

The range masking was used to protect some of the brighter areas during deconvolution? --- or just to block further stretching there?

You went for a parametric psf. It looks to have worked very well. For the deep sky stuff I have been using a dynamically determined psf (averaged from about 30 unsaturated stars in the field). I have wondered whether a dynamic psf would also work for moon and planets etc (as it appears to for planetary nebulae) if --just using the same exposure and everything else --- one were to use a neighbouring star field to determine it? --30ms would be fine with my scope- possibly not at 71 mm?

Local histogram Equalisation looks to have done a great job in maximising local contrast --- did you also try it without. It is one of the many PI tools that I haven't used yet - and thinking about it - maybe should do for planetary nebulae for example?

Final question ....was curves the final step or if not where did it fit in the processs?

Thanks..a lot of questions ...but I am always interested in understanding more of what PI can do and you seem to have got a very good result there.

Tim

Thanks for the comments Tim. The answers to your questions .....

this is a move away from using AS! and Registax to PI for processing lunar images?

It was to investigate alternate ways of processing lunar images. One of my issues has been processing a stack which has a bright limb. Using Registax sharpening will apply across the whole image and start to blow out the bright areas. This limits how much processing can be applied. The use of masks in PixInsight got me thinking and some internet research for 'PixInsight & lunar processing' brought up some useful things to try. I have since found that Registax has a masking capability and will need to research that.

The red filter presumably improves contrast for the bright moon?

The red filter helps by steadying the atmospheric turbulence, particularly with the low elevation - this was captured at 17°. The effect is best seen when the video is played back in SER Player. The red filter also helps with contrast. I usually use a Wratten #25 but happened to have a filter wheel available with LRGB filters installed. For low elevation lunar work red works for me - UV/IR cut & IR685 at higher elevations (I can get up to 60° in Spring & Autumn).

The red filter presumably improves contrast for the bright moon?

The deconvolution was done without a mask.

You went for a parametric psf

I went for the simple approach as this was a first time using deconvolution. Standard Deviation and Iterations were arrived at by trial & error making a judgement on the result. St Dev values tried were 2.5, 2.0 and 1.5 with 1.5 giving the best result. Iterations tried were 10, 20, 25, 30 and I settled on 20. If working on a deep sky image I would be using the dynamic PSF technique as stars would be available. My moon image filled most of the FOV and lunar type short exposures, 0.4ms in this case, would not reveal many stars. I did try Deringing on & off and left it off for this image- it might be needed for an image with larger craters.

Local histogram Equalisation looks to have done a great job in maximising local contrast --- did you also try it without. It is one of the many PI tools that I haven't used yet - and thinking about it - maybe should do for planetary nebulae for example?

Local Histogram Equalisation did bring out the contrasting areas well. Without the inverse mask LHE would have destroyed the image. Without LHE the image was very bland, once applied the image was drastically changed. I think you need to try LHE with the planetary nebula to see the effect.

was curves the final step or if not where did it fit in the processs?

The PI processing order for the Autostakkert generated stack was:

• Apply Deconvolution

• Use Range Selection to generate the mask

• Apply the mask to the image and invert the mask - red areas are protected

• Apply gentle 'S' curve (repeated) to brighten the image

• Apply Local Histogram Equalisation

• Rotate 90°

All values used in the various steps were arrived at by trial and error. As always with PixInsight there are so many variables presented by the various processes. The key to this image was the mask protecting bright areas from processing. The annotated image demonstrates how many craters managed to survive my processing.

The Vixen 81s refractor was used for convenience, as with the reducer the FOV could capture the 78% object. It will be interesting to see what my C8 SCT images are like when processed in the same way.

After this result, I will be going back over some older data to see what improvements can be made.

This session started out as attempt to capture separate R, G & B images. I am having a slight difficulty combining the images and put this down to the fact that I forget to change the tracking rate to lunar on the mount hand controller ;-( Still I have managed to salvage something decent out of the session.

Dave

Thanks very much Dave! A lot of useful learning there that I hope to put into practice - not only for the moon but particularly as Jupiter and Saturn slowly wheel round to culminate at elevations above the treetops and houses.

Had a go at using LHE on some of the M57 images - previously I had been applying a photography program - Photoscape X - to achieve a similar effect post PI. I think that LHE is better and probably more flexible because of the range of parameter adjustments it offers.

Tim

Tim

It will be interesting to see what happens with the planets, lack of features will have some impact on the results.

The big eye openers for me from this exercise have been deconvolution and LHE.

It would be interesting to see some side-by-side comparisons of your M57 with the various processing techniques applied.

Dave

Hi Dave, I have taken up your suggestion and added a few images of interim steps during processing to the thread on M57 and M13. Tim