What to do when Photometric Color Calibration Doesn't...

This image poses a very difficult problem for gradient correction. For example, where are you going to put samples on the third quadrant (bottom left)? I don't see any obvious gradients in the image you have uploaded, although judging them is also quite difficult. The multiscale gradient correction method requires another image acquired with a wider field instrument. In this case you'd need something like a 50 or 35 mm wide-angle lens.
 
This image poses a very difficult problem for gradient correction. For example, where are you going to put samples on the third quadrant (bottom left)? I don't see any obvious gradients in the image you have uploaded, although judging them is also quite difficult. The multiscale gradient correction method requires another image acquired with a wider field instrument. In this case you'd need something like a 50 or 35 mm wide-angle lens.
Yeah - I just read the reference article you cited in your previous post and realized that I don't have a larger image to use for the Multiscale Gradient Correction so that's not going to work here.

So I guess my choices are to do DBE, then PCC, then DBE again, or just skip DBE altogether. What do you think?
 
doing DBE after PCC can upset the color balance unless you tick "normalize" in the target image selection area of DBE. why do you need to do DBE twice?
 
doing DBE after PCC can upset the color balance unless you tick "normalize" in the target image selection area of DBE. why do you need to do DBE twice?
Well, it's either that or SCNR, but if I do DBE prior to PCC then I get the "Hulk green" result that has to be re-corrected.
 
These images can be calibrated without problems with our PhotometricColorCalibration tool. However, default parameters are suboptimal for these wide-field images because they lead to poor sampling of photometry data, so they must be changed to achieve good results.

Take a look at the following screenshots:

Lagoon/Trifid:

Rho Oph:

Default photometry parameters usually work well for most regions of the sky. However, wide field regions covering dense milky way areas pose particularly difficult problems, mainly because of the huge amount of stars and the complexity of the background.

As you can see, I have tweaked photometry parameters in order to ensure good sampling of the APASS catalog. For the Lagoon/Trifid image I have forced a limit magnitude of 12 and a photometric aperture of 8 pixels. With these parameters PCC has performed an absolutely perfect white balance calibration with about 3000 stars from the APASS catalog. The result is excellent, as expected from the white balance function graphs.

For the Rho Oph image I have used a limit magnitude of 10 and a photometric aperture of 6 pixels. These values are more appropriate for a wider field. The result is also very good, although APASS data for the r' filter shows much more dispersion. Despite that, the achieved white balance is excellent IMO.

As for background references, they are not very critical, as you can see in the examples above. Just try to select a relatively free area, such as a dark nebula, and avoid selecting background references too far from the center of the image if you have significant gradients. Use the Statistics tool to define a reasonable upper limit for background neutralization.

Hi @Juan Conejero ! thank you so much for share with me the correct way to perform the PCC. To be honest with you, I didn't expect this photo be so tricky, I learned a lot reading you and all the persons in this Forum. I already perform the PCC on my image and obtain the same result and right now I'm experimenting with other parameters just to understand the process. Thank you everybody! Looks like there is a long way to learn and master PixInsight!

Best Regards,

Rodrigo Quiroga
 
PCC works for me 95% of the time. But there are some images it just refuses to calibrate. I have read this and several other posts and followed all the reasonable suggestions, and still no results. In the end, I calibrate it the old fashion way.

The image is the starfield around OU4. It's bright with lots of stars. The message I get from PCC is the following.

*** Error: Failure to process image file 'C:/Users/Eric/AppData/Local/Temp/PCC_B_JUUJKMJXL5IU.xisf': Could not locate any stars in the image

Of course, this is probably pilot error and I am always open to suggestions. Also, if anyone is interested in trying to solve this image, please let me know and I will upload it.

Thanks in advance for your help.

Eric Coles
 
this happens from time to time and the error message is less than revealing - the problem is that there were no, or not enough stars in the catalog at that RA/DEC to do the photometric analysis. i guess there are some pretty big holes. i think this is the 2nd or 3rd time i've seen someone report that problem.
 
Thanks Rob. That's odd considering the number of stars in the image. Is there any way to match the RA/Dec (21:11:20/59:56:12) to a particular catalog and overcome this issue? I hate to give up.

Eric
 
i think it's just more that there are areas which, for whatever reason, the APASS catalog has poor coverage. check the coverage map here: https://www.aavso.org/apass - you'll see that most of the sky is not covered well and the emphasis is on the milky way. still there are places with holes even in the MW.

i don't know if there are other catalogs that are appropriate for photometry. maybe there are but they would have to be added to PCC...
 
I'm experiencing another problem trying to plate-solve in PCC. Any single images resolves quickly and accurately, but the stacked image fails every time. There are some clouds near the bottom of all the images but that doesn't seem to be a problem when plate-solving any individual image.

I posted the first and last RAW images along with the stacked & cropped image (which fails) here.
 
the image solved OK in PCC using the following settings. i do wonder if it is a good idea to leave the bottom part in since the stars down there are probably discolored by the clouds and may throw off the PCC result.

to get the center coordinates i cropped a small area from the middle and sent it to nova.astrometry.net. also had to load one of the CR2s to discover the camera model and lens FL and then google the pixel size.

rob
Screen Shot 2020-08-16 at 5.59.08 PM.png
 
Rob,

My apologies - I should have provided the RA/Dec info and the pixel size for you. The focal length and time can both be obtained from the image, of course.

The clouds are a definite thorn in my side. My intention was to clean up this image and use it as a background, so the foreground would cover up the cloudy area. But I couldn't get past the PCC step. I wonder if it would be possible to add a feature to PCC to allow the user to select a region to use for the plate-solving and color calibration, and then use that profile to apply to the entire picture. Would that work around the cloud issue? If the cloudy area gets messed up, it doesn't really matter because it's already messed up and probably not usable in the final image anyway. But for sizing, layers, and blending, it's helpful to retain the entire image.

Anyway, I entered these values and reran PCC, and it still failed. Here is a screenshot of the settings I used which I believe are the exact same as the ones you cited. I also attached the log file for reference. Where do I go astray?

1597678512080.png
 

Attachments

  • Log.txt
    10.3 KB · Views: 52
strangely, i ran it a 2nd time after posting that and it failed to solve. i needed to adjust the BN parameter to not be so high (i think it should be something like 0.0036).

i changed the Alignment Device (method) to Polygon Matching and then it solved right away. i'm not sure what's going on with this but i suppose there is a little bit of randomness in how the solver picks a subset of stars to match. anyway, give that a try and see what happens.

in theory what you can do is to crop your image and run PCC on the cropped image, then use pixelmath to multiply the channels of the uncropped image by the factors given on the console at the end. i think in order to avoid confusion with the background neutralization step you can do that on both cropped and uncropped images before running PCC using the BackgroundNeutralization process and then turn it off in PCC.

rob
 
Thanks, Rob. Changing those 2 parameters enabled it to plate-solve (although it's still downloading the Vizier data and taking forever).

Good advice on the workaround suggestion as well. It still may be a useful feature to add to PCC at some point, though.
 
I'm running into the problems noted originally in this thread. I have an image of Sh2-101 taken with a Tak106 and QSI683. Its 530 FL with 5.4 micron pixels. The image plate solves without issue both with the image solver script and the PCC.

However once the PCC process begins looking at stars for color calibration it reports a certain number of stars in invalid positions. It goes through various iterations and then fails. I have tried changing the limiting magnitude as well as the aperture to 6 and 8 as in Juan's previous posts. I've increased the wavelets for noise detection. I cannot get this image to color calibrate. Any ideas? Thanks
Carlos

Screen Shot 2020-08-20 at 10.23.59 AM.png
 
PCC works for me 95% of the time. But there are some images it just refuses to calibrate.....

*** Error: Failure to process image file 'C:/Users/Eric/AppData/Local/Temp/PCC_B_JUUJKMJXL5IU.xisf': Could not locate any stars in the image
this happens from time to time and the error message is less than revealing - the problem is that there were no, or not enough stars in the catalog at that RA/DEC to do the photometric analysis. i guess there are some pretty big holes. i think this is the 2nd or 3rd time i've seen someone report that problem.

So just to be clear -- the above error message is not about stars in MY image, but that the catalog(s) in use do not have enough coverage?

I've been changing magnitudes, log sensitivity trying to detect more stars, but this is saying the issue is where I was pointing not the image?

Is there any point in changing catalogs or database server or something, or is the conclusion from that error "can't be done stop trying"?

PS. My shot is centered on M101 at 2800mm cropped square around the galaxy arms, the star search finds 249 valid, 20 rejected.
 
So just to be clear -- the above error message is not about stars in MY image, but that the catalog(s) in use do not have enough coverage?

I've been changing magnitudes, log sensitivity trying to detect more stars, but this is saying the issue is where I was pointing not the image?

Is there any point in changing catalogs or database server or something, or is the conclusion from that error "can't be done stop trying"?

with some testing i just did now, i'm not convinced of what i said above. i cropped one of my M101 images to approximate your FOV and PCC still worked, so i don't think there's a lack of stars in the catalog in the vicinity of M101, even with such a small FOV. there are definitely very few (like 30), so there is a chance that PI is not able to find those same 30 stars in your image rather than the catalog being empty.

can you post your .xisf so i can try it?

rob
 
i don't think there's a lack of stars in the catalog in the vicinity of M101, even with such a small FOV.
The FoV is 29' 23.4" x 28' 13.2" (centre RA: 14 03 17.180 Dec: +54 22 05.78 = RA: 210.821583333 deg DEC: +54.368272222 deg). Using the APASS process with APASS DR10, with a search radius of 15', I get:
Search results:​
Total sources .............. 0​
Extending the search to a 30' radius, I get:
Search results:​
Total sources .............. 5​
So "not enough sources" looks probable. The APASS catalog is heavily biased to the galactic plane (to be fair, there are a lot more stars there), while M101 is about as far from the galactic plane as it is possible to get.
(due to dither)
I guess you mean "due to drizzle".
 
Back
Top