0.00
#...
·
 |
|---|
|
I am looking for help with light areas, generally oval shaped, in my narrow band subs, primarily with the Oiii and Sii filters. I recently imaged the Horse Head nebula, but noticed several light oval areas in the Oxygen and Sulfur bands. I used the same filters in October, and didn't notice these light areas then, so something changed. These light ovals are not noticeable using the Ha band. I have redone my flat images for my narrow band filters, but that did not resolve the problem. I associate dust motes as dark areas, but have no idea what would cause light areas. I cannot understand what would cause areas of my narrow band filters to be lighter than the rest of the image. Does anyone have any idea about what would cause this and how to resolve it? As I said, I took new flat calibration images but the light parts remain. Thanks, Jim |
|
7.76
#...
·
|
|---|
| Maybe posting an example? |
0.00
#...
·
|
|---|
| I would go so far as to say that these ovals are due to an overcorrection of the flats in the calibration process, but I would need to see an example |
|
0.00
#...
·
|
|---|
|
Santiago, I am not familiar with "over correction of the flats." I used NINA's flatmaster to create flats and used Pixinsight BWPP to use them to calibrate my subs. The histogram for the flats was approximately in the middle. Below are samples of my images. The first attached is a single, raw, uncalibrated, subframe using my Oiii filter, and below is a result of a stack of 53, 10 minute, subframes. The light ovals are apparent in the upper middle right of the raw frame, and more prominent in the lower left of the stacked image. I took new flat frames using NINA's flatwizard, hoping that would resolve the defect, but they were unaffected. Someone else suggested it was "over correction" from the flats, but I an not familiar with that term. The flats were taken with the histogram at approximately 45% to 50%. The light oval defect is barely noticeable in my Sulfur filter images, and not noticeable at all in my Hydrogen images. In case anyone is interested, the lower images is my Oxygen filter master flat. The attached files are JPEG images with a simple ScreenTransferFunction stretch using Pixinsight, and the Pixinsight images converted to JPEG and reduced by 75%. I did not notice this defect in images taken in October (between the holidays and weather I took none in November or December), and am at a loss as to what changed, and more importantly, how to correct it. Any help would be appreciated. Regards, Jim    |
|
7.76
#...
·
·
1
like
|
|---|
| They are just halos from the bright stars in (and off) the field. This tend to happen with some brand of narrowband filters. And refractors. You seem to have a lot of halos on-stars too, so it is worth considering a change of filters. In lieu of this some of those "phantom" halos can be corrected (removed) in post-processing. |
|
0.00
#...
·
|
|---|
|
Andrea, Thanks for your thoughts. You may be correct about the halos. I have been imaging for about 12 years with an old 8" Newtonian, an old SLR camera, and manually changed narrow band filters. I recently upgraded to a Sky Watcher Esprit 100, filter wheel, OAG and QHY 268m, and I'm still learning my new rig. So far I am pleased with the upgrade, but the halos are new to me. I understand that Oiii filters are notorious for halos, so that may be part of my problem. A tighter bandwidth on the filters might help, but that's for another time. I am using 2" mounted Baader filters, and mostly see the halos with the Oxygen filter. The halos are barely noticeable with the Sulfur band, and not noticeable with the hydrogen filter. Do you think reflections might from the filter mount rings might be the culprit? Or maybe it is just the nature of the Esprit 100 with its proprietary field flattener, and bright stars. Also, I notice that the halos do not change with dithering. Not sure whether that is relevant to diagnosing the problem. Any suggestions about how to resolve the halos in post-processing? The data I have looks pretty good except for the halos. Thanks, Jim |
|
7.76
#...
·
|
|---|
|
Hi Jim, Alnitak (and her sisters) are notorious for these and related issues as in if there is any chance to create an artifact then it will happen. Cure, there is none. I have seen them even with 3 nm passbands filters (and mostly refractors) so it is a combination of factors. You can remove them to an extent depending on the particulars of the situation which in your case are highly favourable. Sadly, it needs been done on all the offending frames as it cannot be automated and it needs PI. If you have both the time and the inclination I shall endevour to describe the procedure for you. |
0.00
#...
·
·
1
like
|
|---|
| You might get lucky and maybe all you need to do is turn your filters around. Sometimes that can make ALL the difference with internal reflections. |
|
0.00
#...
·
|
|---|
|
Andrea, Thanks for your response. Nice to think the problem is not entirely me, but then... I have Pixinsight, and a rudimentary understanding of of the software. I always look forward to learning new tricks. What do you propose to "cure" the reflections in my images? I devoted a fair amount of time in my limited clear skies capturing the images, and in a rough compilation the noise level seems tolerable. I think I can do something with them if I can fix the reflections. I suppose I could crop them out, but I was hoping to do a wide field view if possible. If you can help me salvage the reflection images any help would be appreciated. Thanks, Jim |
|
0.00
#...
·
|
|---|
|
drmikevt, Thank your for your suggestion. However, my filters are mounted in 2" threaded rings, threaded into my filter wheel, so I don't think there I can flip them. i would try if I could, but my filter wheel won't let me. My filter wheel would permit me to flip them if they were not mounted. Something to consider if I upgrade. Relatedly, I have had these filters for a while and am considering upgrading them. My existing filters have a general 8 nm band width, but am considering upgrading to something like a 3 nm bandwidth. What filters would you suggest? Again, thanks for the suggestion. Jim |
|
7.76
#...
·
|
|---|
Jim: It would be best if you could make available a pre-processed single frame, (linear but otherwise fully calibrated) so that I could illustrate the process in step-by-step fashion. |
|
0.00
#...
·
|
|---|
|
Andrea, I have the calibrated files, they are 102mb each in xisf format, but astrobin doesn't support downloading that format or file size How should I send it to you? Jim |
|
7.76
#...
·
|
|---|
| I guess Google drive or similar service. I reckon it is free. |
|
0.00
#...
·
|
|---|
Jim: All I can say is that I love my Chroma filters. I have the 3nm set and get zero reflections or star bloat/haze/glow. Are they expensive? Sure, but you only need to buy them once. |
|
0.00
#...
·
|
|---|
|
Andrea, I uploaded the files to Dropbox, but don't know how to get the link to download it to you. I'm new at this, but I think I need to send you a link. Jim |
|
7.76
#...
·
|
|---|
|
Jim, That's right. You need to share a folder and send the link to that folder so that I can download it. |
|
7.76
#...
·
|
|---|
|
Hi Jim, It works. I'll post here for the procedure to follow. Cheers Andrea |
|
7.76
#...
·
·
1
like
|
|---|
|
Part 1. Removal of halos and shadows fro linear image (essentially, off axis reflections). 1. Remove stars from each single frame affected (after all the usual image calibration is done, that is before any such steps as de-bayering or optimization that precede registration are carried out). This is best done with starnet++ V2 latest version NOT SXT!. Remember to select linear data in the options, too. The result is shown below:  2. As an example we are interested in removing the off-axis halo highlighted by the preview box as shown above. Since the data is linear (no stretching is applied) and the high likelihood that the off-axis ghost reflection will not fall in the same location when registration is done we need to remove it from the stack of light frames for all of them. 3. Create a mask that contains only the halo to be removed. This is easier said than done when dealing with unstretched images as you would have to set the threshold very very carefully, as shown below:  4. As shown the results cannot immediately used as mask as you need to isolate only the patch that contains the reflection(s) you want to remove. First step is to binarize the image setting a threshold that captures the halo. After that you need to blot out all the other areas with the lightest (if not white) area you have in the frame (using the clone stamp). Once all the other areas have been eliminated you end up with something like this:  5. Now the task is to work out a way to smooth out and remove the dusting of black pixels remaining, which can be achieved by applying a blur of 2 pixels, followed by another binarization and finally applying a morphological transformation, as shown below:  You can check whether the mask overlays the reflection when you iterate through the MT operations. Once you are happy with the way it masks the affetced spot move on the the core of the operation. 6. Apply the inverted mask created in the step above and create a preview of the zone you'll be operating on and launch the MultiscaleMedianTransfrom Process, as shown below. Select as many layer as need, normally at least 8 and possibly more depending on the image and the amount of gray levels you have. In this case 9 layers ought to be enough:  7. Unselect one by one from the layer n.1 to the layer n.9 and see how that affects the preview. Usually the last few would completely delete everything and the first would smooth out the high frequency pattern in the affected zone. You then back off and add back layer after layer until the pattern is removed and you're left with a similar noise pattern of the background nearby:   8. Now back off from the preview and apply the MMT process to the whole image:  And then add the stars back:  Very faint halos away from any significant structure as the weak one shown above can be dealt with, in most cases, by blotting it out of existence cloning a nearby background areas (sans stars, obviously). Part 2 to follow showing how to deal with star halos... |
|
7.76
#...
·
·
1
like
|
|---|
|
Part 2. Removal of star halos. This procedure needs apply to the stretched final stacked up image (channels or RGB). It follows the same blueprint as discussed in Part 1 but with the caveat that masking isn't a single operation but a sequence of such as the halo might not be just one but several of different intensities so the operation of halo removal is a staged operation and only applicable whenever there is clear distinction between it and the background, that is it is a proper halo (aureole) rather than a diffuse glow. What is singularly difficult is the preservation of a star core will suppressing or reducing the brightness of the halo. In this ecample the first step is to blunt (but not entirely remove) the main, brighter halo of Alnitak. The stretched (single frame) image is shown below:  The mask of the Alnitak halo is shown here below:  This mask encompasses both halos but leaves the star "core" free so that at the end you are still left with Alnitak, not a hole (black or otherwise grey). The MMT setting for the halo reduction are shown following here:  And the final image:   The 2nd iteration will be much simpler as now the star removal operation will actually remove the star leaving a uniform smooth halo that could be easily reduced with a further masking and application of MMT, as can be seen here (this is just the starless version, no actual halo removal has been carried out yet):  |
|
0.00
#...
·
|
|---|
|
Andrea, Thank you. I have saved your process and will give it a try. Regards, Jim |
0.00
#...
·
|
|---|
| (deleted) |
1.81
#...
·
|
|---|
andrea tasselli: Alnitak (النطاق, arab, m.), is a brother  |
