Compiling ADU values across batches of images

[BrendanC]

This might be a slightly oddddd question, but I've looked online and can't find an answer. Basically: is there a standalone utility, or feature within another program, that can analyse the ADU values across a set of images? That is, as a batch, to avoid having to do this manually? So, I just point it at the set of images, it goes through them, and gives me the ADU for each image? I've looked at PI's Blink feature, and the quality metrics in software like DSS and ASTAP, but none of them include ADU.

[ONIKKINEN]

Siril can sort of do this, but you'll have to jump through some hoops to get the values. You can import a sequence and register it to align the stars, then you can draw a selection around some medium brightness star in an area of no nebulosity and select the "PSF for sequence" option with right click.

Normally you will have 32-bit files at this point since you are registering frames that have been presumably calibrated, so what the graph tells you will be in 32-bit values so not exactly the simple ADU number you were looking for. But if you do this with 16-bit files and set Siril to 16-bit mode, you will of course get a 16-bit output on the graph and now it is more or less a straight up ADU value (for the background), tested this and it reports the background value in a 16-bit value.

 

Did you have some specific reason you wanted to have a simple ADU readout of all the frames? This method is not really ideal as it has you do all the extra work, but you can get a graph out that tells you what the background values in each sub were.

[BrendanC]

Thanks - looks quite convoluted but I'll see what I can do.

The reason is that I recently realised I could in theory analyse all my previous data by exporting the file details into Excel, using the time/date to calculate where it was in the sky, and then match that to the ADU value for that specific image. That way, I could get a rough idea of the brightness of different areas of the sky where I shoot, and adjust my exposures accordingly. One side has a lot of streetlamps, the other doesn't, so I'd expect to see the ADU value be higher on the light polluted side, which means in future I could reduce my exposure time on that side to compensate.

Or, to put it another way: I just like cool charts of lots of numbers because I might be able to see interesting patterns in them that may - or may not - help me with my astrophotography.

[vlaiv]

39 minutes ago, BrendanC said:

and gives me the ADU for each image?

What would be ADU value for each image?

 

[Seelive]

Years ago I couldn't find anything to do it either so created my own Excel spreadsheet using VBA to do it but it currently only reads PIC image files (IRIS images), rather than TIF or FIT image files. It was designed to calculate and save the statistics of each image (max, min, mean, median, standard deviation, background noise and histogram peak ADU values). Perhaps there are more versatile astro-imaging packages available these days.

[BrendanC]

Ah, I wondered when @vlaiv would turn up!

The average ADU? Like for example the figure APT shows on its Pixel Aid tool.

You're going to tell me I'm doing something ridiculous. I can tell.

[BrendanC]

3 minutes ago, Seelive said:

Years ago I couldn't find anything to do it either so created my own Excel spreadsheet using VBA to do it but it currently only reads PIC image files (IRIS images), rather than TIF or FIT image files. It was designed to calculate and save the statistics of each image (max, min, mean, median, standard deviation, background noise and histogram peak ADU values). Perhaps there are more versatile astro-imaging packages available these days.

Thanks, and I'm starting to wonder whether I might use my JitBit macro recorder to do something similar.

[vlaiv]

ImageJ can do all sorts of nice features.

If you want to measure stack in one go - just select a region in stack (stack being sequence of images in ImageJ terminology) and perform command.

This puts results in table that you can export or copy/paste in spreadsheet for further use.

 

[ONIKKINEN]

5 minutes ago, BrendanC said:

Thanks - looks quite convoluted but I'll see what I can do.

The reason is that I recently realised I could in theory analyse all my previous data by exporting the file details into Excel, using the time/date to calculate where it was in the sky, and then match that to the ADU value for that specific image. That way, I could get a rough idea of the brightness of different areas of the sky where I shoot, and adjust my exposures accordingly. One side has a lot of streetlamps, the other doesn't, so I'd expect to see the ADU value be higher on the light polluted side, which means in future I could reduce my exposure time on that side to compensate.

Or, to put it another way: I just like cool charts of lots of numbers because I might be able to see interesting patterns in them that may - or may not - help me with my astrophotography.

Well in that case you might find the plot function in Siril very useful. There are FWHM, Roundness, Amplitude, Magnitude, Background, and SNR readouts for the plot. The relative magnitude, background, and SNR tabs will easily show you which subs were duds and which were not. You can select these subs from the plot to exclude them from further processing, so when stacking these are not put into the stack. I use this all the time to remove the statistical outliers from stacking.

[vlaiv]

1 minute ago, BrendanC said:

You're going to tell me I'm doing something ridiculous. I can tell.

No

You can go about this in several different ways.

Average will be comparable only on single target and only if you don't shift a lot with your FOV (regular dither is ok).

If you want to measure background - probably better way would be to use median.

Alternative would be to take stack with removed background (just make sure you use sensible stacking algorithm that won't do weird stuff like scale output in 0-1 range, but leave ADUs as they are) and then subtract from each sub and then measure average.

For this to work - you need to have your subs aligned and cropped to have some data in all subs (trim edges that would otherwise cause stacking artifacts).

In any case ImageJ should be your go to tool

[Aramcheck]

1 hour ago, BrendanC said:

 I've looked at PI's Blink feature, and the quality metrics in software like DSS and ASTAP, but none of them include ADU.

PI's SubFrameSelector will give you the info for a batch of files, which you can then download to csv. In the Data Unit drop down you can select between measuring the subframes in Electrons / Data numbers or Normalised (0-1).

Cheers
Ivor

[BrendanC]

Nice one, thank you!

OK, so I'm feeling my way through this, and I don't know which metrics to select in the drop-down menus (there are two?)

I can't find Electrons / Data numbers or Normalised (0-1).

What two(?) would I select just to get an average ADU?

Edited November 16, 2022 by BrendanC