Flip Inspector Results
Stacking program, astrometric (plate) solver, fits file viewer
Brought to you by:
han59
Menu
▾
▴
I may be doing something simply wrong here, but I use capture programs and with a RASA, ASTAP requires me to flip it vertically in order to match what comes in via most capture software?
No big deal, I will flip it when I do a single image tilt or curvature measurement for example and I can clearly see the stars aligning with what all my capturing software such as NINA use. Analysis matches this way.
But I also do a focus curve and will run through ASTAP -> Stack -> Inspector and will then copy this data to spreadsheets. The problem I am having here is that the images are still flipped vertically. My top left is actual sector HFD 11 (bottom left in ASTAP inspector) and my bottom left is actually HFD 13 (top left in ASTAP inspector).... regardless if I have the flip vertical setting on or off.
2 images show pixinsight star positions (same as NINA and other capture software), while astap shows flipped vertically. not a big deal, but it does mess with inspector location within multi-focus runs. I can handle in spreadsheet most likely, but be nice if it matched up with setting under Tools menu? maybe I"m missing something that does this?
Also, is there a reason it's flipped in ASTAP but not across any other capture or raw image files?
Just to follow up, here is what we are doing with spreadsheet data, we feed it into a ton of raw calculations, graphs, etc... and just realized that even though we flipped it vertically to match and align in the single image analysis of ASTAP, the stack -> inspector section is still inverted.
We can make it work in spreadsheet by building it in to flip, but would be nice to cut it off at the head of the analysis which is ASTAP... if possible.
Hi Josh,
Traditionally pixel 1,1 of the FITS is displayed left bottom. This convention is not followed by every program. But you can flip the image under view.
The next problem is the wild west around the definition which pixel has coordinates 1,1. Normally that is the first pixel inside the file. ASTAP is following that convention.
In ASTAP for the tab inspection the reported sector HFD11 start at pixel 1,1. A vertical flip doesn't change the situation. If the image is vertical flipped then pixel 1,1 is top left and so is sector HFD11. So a vertical flip in menu view doesn't change the report.
So ASTAP is following the FITS standard and conventions. I don't think PI is following the conventions. So how could this be improved?
Han
First off, you have some great software here.... and as I do software, I want to congratulate you on it. We build the ASG Photon Cage which we use these tools heavily as a tilt device for astro cameras...
so here is another image of the 2 single image analysis, which you can see flipped and not flipped that corner with 3.63 is what I am tracking here. They work great.
But if we do a stack inspector, HFD 13 is always the 3.6 corner... even if we flip it, these results and the corresponding labels on here show it as top left, so it never matches the flipped single analysis image.
if we take this data to our spreadsheets, HFD13 is no longer top left, but bottom left with the data... correct? or maybe I"m missing something? just icons are wrong on the data? but I"m not seeing data change at all in stack inspector?
it 'feels' like the stack inspector data should either change icon positions here in columns or values should change and correspond to the flipped value.
could just be my mind playing with me... lol
Your two image flipped or not flipped show correctly the visual flipping. But note that pixel coordinates are following the flipping. So when view is vertical flipped, coordinate [1,1] is at left top. Normasl it is at left bottom
For the tab inspection the view flip has no influence. The nine areas are locked to pixel coordinates so to the sensor itself. Why should it be different? Pixel [1,1] is in HFD11 and the pixel with coordinates [width,height] is in HFD33 always. This allows comparison with other programs/viewers. The pixel coordinates in the FITS file are standardized. According the FITS standard 4.0:
This is also essential for the WCS. astrometric solution. The only program I know which doesn't always follows it is PI. I have brought their discrepancy to their attention but they are not interested in fixing it.
ASG Photon Cage looks impressive. But a high price tag.
Han
makes sense regarding pixels being in the correct spot always. I think the confusion then is in the icon displayed in the inspector... that is all I was saying, the single image flip lets you say 'this is top left'..... this lets you match NINA, Pixinsight, etc... as you said anything goes. So I had been aligning images to those platforms, which requires a flip vertically. Single analysis aligns, top left matches top left on others easy enough.
At this point, it shows in single frame analysis that bottom left is the issue... So I dump a focus curve into inspector and it shows HFD13 (top left) is the issue. At least according to the little icons? I think that is where I may be getting backwards is following the icon locations?
I agree the descriptions like "bottom left " are not ideal. But how else to describe them or skip the textual descriptions?
There is an English saying "going south" meaning failure. But what has to do with direction south?
20:55:38 area 1,1 (Bottom Left) Focus=5932 a=7.45893 b=896.97690 offset=-13 error=0.01990 iteration cycles=6
20:55:38 area 2,1 (Bottom Middle) Focus=5955 a=7.16258 b=856.77334 offset=11 error=0.03183 iteration cycles=9
20:55:38 area 3,1 (Bottom Right) Focus=5889 a=6.85708 b=745.60471 offset=-56 error=0.04805 iteration cycles=10
20:55:38 area 1,2 (Middle left) Focus=5968 a=6.75846 b=830.93328 offset=23 error=0.01539 iteration cycles=9
20:55:38 area 3,2 (Middle Right) Focus=5909 a=6.42768 b=733.58765 offset=-35 error=0.03400 iteration cycles=9
20:55:38 area 1,3 (Top left) Focus=5958 a=7.13608 b=819.71249 offset=14 error=0.02364 iteration cycles=10
20:55:38 area 2,3 (Top Middle) Focus=5884 a=7.12359 b=850.11261 offset=-60 error=0.03203 iteration cycles=9
20:55:38 area 3,3 (Top Right) Focus=5916 a=6.71384 b=754.00398 offset=-28 error=0.04271 iteration cycles=7
oh I agree, it's probably a endless rabbit hole trying to accommodate this... I will need to just build something into my spreadsheet. Here is the little icon I was following.