Mr.han

live stacking for EAA, please check Sharp CAP for windows app.
A slider that adjusts the brightness and color of the stacked images can be used conveniently for viewing.

wow these are impressive results indeed!! Really looking forward to integrating this feature as well with KStars, just after I finish the ASTAP astrometry integration. Great work Han!

I will do. But sliders are already available.

I'm trying to modifying an application which was designed to stack images from a stationary telescope with guiding. No guiding is no problems as long you keep the exposures short. Currently there is a concept problem that the application doesn't know what's happening with the telescope. When your continuous imaging and move the telescope you get something like this:




This will stack since a part of the exposure is good but somewhere halfway the operator decided to slew. But this will spoil the stack. Looking into the FITS header of CCDCiel there is no keyword I could use to identify the start slew halfway the exposure. So you either have to 1) suspend imaging automatic after a slew command or 2) add information to the header indicating the mount has moved or 3) ASTAP reads out the mount position. 4) stop manually the exposures

This is something for Jasem (Kstars) and Patrick (CCDCiel) to advice. Step 3 is a big step, so option 1) or 2) look the better option. Option 1) looks to me the best option since imaging and slewing doesn't make sense and any image will be worthless. Maybe this is already available in Kstars, I'm not familiar with it.

Han

Fixed some run time problem with ASTAP release v0.9.282. This version processes all my images from last night, good or bad.

Han

Han, this look very good!

You can detect if the telescope as moved with the RA and DEC FITS keyword.
But it is necessary to compare with some tolerance because the mount driver can update the coordinates after pulse guide command.

This is probably the most universal way to deal with the problem. For example many people doing EAA use the telescope handpad to move the telescope and the imaging application is not aware of that.
But the most reliable is a Start/Stop button in the stacking application, so it not stack until the user finish to center the object and sometime focus manually.

Patrick

A 'simpel' levels with histogram is very convenient for adjusting the stacked image.

I already compar the FITS header position with 0.2 degrees accuracy. Smaller offsets will be stacked without problems assuming the image is covering an area of 0.5 degrees or larger

I just did a test with CCDCiel imaging and slewing at the same time . If the mount is moved during exposure, the FITS file is written with the latest position in the header. That's good. So it means that I have to skip an image with a new position. Currently in ASTAP live stacking an image with a new position is used as the first and reference image. That's why it goes wrong. I will fix that in version 0.9.283

Hopefully Ekos does the same (writing the actual mount position at the moment of saving)

That's already available.

To clarify it for some users. Live stacking will be processing files using a dedicated directory. (Like the Windows application "deepSky stacker live") There is no link to a camera or mount. Actual imaging should be done by Ekos, CCDciel or any other application. The ASTAP live stacking program will just load and stack the available images in the dedicated directory. If all images are processed it will wait for more. Exposures should be 2 second or longer. It is not suitable for planetary imaging. If a new position is detected in the FITS header, it will start a new stack. To identify processed and unprocessed files, the processed file will get the .fts extension. Images will be stacked correctly independent of shift, rotation or flipped. The stacked image will be displayed after each addition and the image quality with the number of images added. It will work for unguided images as long the exposure times are short. The ASTAP viewer will work effectively like a slow scan video camera.

Han

Some personal observations - sorry but it was a Windows version of ASTAP running on my test rig - an older AMD A6 8GB Win 10 64bit with SSD. I used a DSLR and no Scope set up. Old stored images were used for he live stack. Nothing else out of the normal being used except MQTT broker on Windows!

1. Too complicated IMHO - by that I mean you just need to set the folder where the images are going to be stored and allow the user to end the session and change the stacking source does matter that the scope has slewed in the mean time as I would expect the Live stacking to "Lag" behind by a number of images anyway - ok more difficult when using automated sequence. Plus Astrotoaster allows a fair amount of simple changes,in real time, to the stack (well the viewers displayed image) just by using simple sliders on various parameters via a simple cild window called "Col Adj" - something like that would be great in ASTAP

2. On my test rig ,which runs DSS Live and Astroaster ok ,I was unable,once live stacking started, to do anything within ASTAP as it kept displaying the dreaded revolving circle and "no responding". Task Manager showed very high processor usage but no more than either DSS or Astrotoaster - figure shown was around 34% with only 2.7gb out of 7gb of memory used.

3. I could not find a way to just monitor RAW (CR2 files) as per DSS etc and let ASTAP do the conversion if required to FITS. I know ASTAP will convert RAW files to FITS for normal stacking,

4. Personally I would prefer the Live Stacking to be a separate application and NOT part of ASTAP - simple and clean which both DSS Live and Astrotoaster are ,even though they use DSS in the backgound - but I never need to start or see DSS !

5. The ability to reject some images in the Live Stacking list and restack would be nice.

As I say this is my EAA personal observations and know it will be used in other ways ,so it is a blinkered view

Plus what is the hardware testing set up beings used by others - RPI ? or other - obviously running ASTAP on new faster kit and maybe Linux OS would my problem as noted by point (2) but would be interested on what others are using during this test / development phase.

Thank you for a wonderful job.

Many people running EAA with SharpCap enjoy using an automated deployment platform such as AZGTI.

After introducing the target object while watching the live view, press the stacking button to start stacking.

Stacked images are adjusted manually (in color, brightness, etc.) in real time with a slider, etc., and they are having fun.

When moving the target, press the stop button to stop the stack before moving.

Q. Can the following items be realized with this stacking function?

1. Stacking using live view video.

2. Adjust brightness and color in real time during stacking.

I checked the updated app, but it looked like a stacking method to load a saved still image. (It is similar to the function of AstroToaster.)

It will be even better if the above 1.2 features are also added.

Unfortunately video is not possible. ASTAP live stacking is similar in function as AstroToaster&DSS. There are no functions to read cameras and there are no plans to add them. ASTAP could be used as an addition to existing programs like Ekos & CCDCiel.

Adjustment of brightness and colour is possible during stacking. See sliders in the top of viewer. However if used in auto mode [low,medium, high, gamma setting] there is no need to change them. To save the result you could pause/stop the stacking and save the viewer image as FITS/PNG/TIFF for further processing. You can also live zoom in and focus on existing details.

Version 0.9.283 is now released. If no more bugs are found, this should be the version for the next days. Then I will compile for Raspberry Pi and hopefully it is fast enough. The MacOS version I will do tonight.

The only remaining question for Ekos is if the FITS file are saved with the lastest ra,dec readings. This is required for the correct functionality. As soon the ra, dec changes that image is skipped. The next image is used as reference for the new stack.

Han

Stash,

In the first versions, I made a coding mistake with background detection and it was very CPU intensive trying process noise as stars. The latest version should be much faster. You could just test it on a directory of images and see the processing time. I have still have hope it will work on the Pi4.

I haven't added the RAW files functionality yet to the live stacking. That's easy to add but it uses DCRAW for conversion and this could add some delay. If DSLR is used in combination with Ekos are the results saved as FITS or RAW?

Han

Hans ,

Thanks for your efforts - goes without saying really

Thanks for the reply and I note the "error" which is only to be expected in testing - I will download the newer version (using stand alone version of ASTAP not the installer versionanyway)/

The added delay of using DCRAW could / would be removed if you processed images directly in RAW mode as EKOS can provide either if my memory serves me ! I think CCDCIEL converts or receives FIT anyway so CCDCEIL would need to be changed to allow RAW DSLR files - I am using Indiserver on Linux with DSLR attached but running CCDCIEL on Windows so that I can do live stacking via AT - works very well !
CCDciel ,which I have started using live ,does not work too well on RPI as PC says on his site - not sure how well it works on RPI4 - so that really means off OTA processing IMO.

Of course being Indi ,ASTAP Live stacking could be run on a separate RPI (or other hardware) by itself if CPU/Memory is a problem or interrupts workflows (older RPI) it only needs to access the files where Ekos stores images. I have done this before with AT picking up images from Ekos across a wired network (wireless too slow) or using a File sync application.

Still to get a "live Stacking" application in Linux is a big plus.

I will let you know how my testing goes