On the 14th of November 2020, during the Mars apparition and after opposition/ closest approach, I did an extensive series of runs on Mars. It was so interesting to see the development of a regional dust storm, including dramatic flow down the large canyon, the Vallis Marineris as well as bands of blue-white cloud at the terminator. The seeing at the outset appeared to be really good, but it declined somewhat throughout the run.
I thought that this would make a great image, however, I was disappointed to see what appeared to be a ghost image of the dust in the canyon towards the north (up).
Revision B is what I came up with at the time. and here is the original image I posted:
https://www.astrobin.com/a93zjh/The ghost image of the dust in the Vallis Marineris is obtrusive and really spoiled the result. It didn’t help that I ended up over-sharpening it as you can see from the geometric patterns, where edges and boundaries have been over accentuated. The image looks a bit like a cartoon and that is a no no.
I have always wanted to go back and fix it or at least to attempt to. Given the recent inclement weather I’ve had that opportunity.
This image is the result. I managed to lose the ghost artefact…..I’ll explain how shortly. Here the image is sharpened correctly, at least as far as I could go without producing artefacts. Now it looks like a photograph, not a cartoon. I hope you agree.
The ghost image of the dusty Vallis Marineris is, I believe, the Mars edge rind effect but impacting a feature on the body of the planet. That is unusual as it normally manifests at the limb where there is a very sudden transition from light to dark. In fact there was an edge rind effect there but the derotation of 3 images over 40 mins in WinJUPOS largely eliminated it. I have also seen it above the polar ice cap. The edge rind effect is a real diffraction effect, so optical in nature. It is therefore very difficult to eliminate and tends to occur when there is a sharp transition. I think in this case the very bright dust with a sharp edge caused by the walls of the canyon, with good seeing has resulted in the effect. It is of course worsened by sharpening which accentuates edges and was made worse again by my over-sharpening. Of course, I could have taken the image into Photoshop and attempted to darken the offending ghost image, using some of the tools available, but that would have been an unsavoury manipulation of the image. However, I noticed that an image I produced from the final 3 runs that night, as the seeing fell away, did not have the artefact. Clearly as the seeing deteriorated the sharpness of the boundary decreased and the diffraction effect was to all intents and purposes absent and therefore not present to be exacerbated during sharpening.
This image is shown in Revision C and you can see the quality of the image is relatively poor, but the area north of the Vallis Marineris is quite well resolved and without the artefact.
Now WinJUPOS has a feature whereby with an image or images, you can select the Universal Time to which they will be derotated and integrated. I used this feature to effectively derotate the later image (13:03UT) to the timestamp of the primary image (11:07 UT). The result is shown in Revision D with some steps to match the brightness and colour of the image I was trying to rectify. Of course the limb gets really messed up when you do this, but I was not concerned about that.
I put this image in a layer under the main image in Photoshop and after some care to align the features, I was able to use a mask to paint away the artefact revealing the unaffected terrain underneath. With a bit of feathering the result is almost perfect. What I like about this is that I was able to adapt real imagery of the area to effect the rectification.
This was rescuing victory from the jaws of defeat and I am really pleased to have this image.
Work for cloudy days and cloudy nights
Comments