Post WBPP Master Image Contains Frame Edges of Integrating Multi Session Images from Alignment

Professor

I have been wrestling with an issue that I have not been able to find a resolution/explanation in the multiple forums and have not come across in Adam's tutorials. The problem is of frame borders appearing in final integrated image of the Rosette of two sessions. First session is 60s at 150 gain and 2nd session is 120s at 150 gain. Each session has its own flats. The darks are from a library of 60s and 120s respectively. Bias files are also from a library common for both sets of darks all at 150 gain.  Used keyword grouping for pre-calibration into one image. The two sets are integrated and the Rosette itself looks good as well as the stars. However, there appears to be straight lines at 45 deg angle running through the upper left and lower right of the Rosette with perpendicular lines on the other sides of the Rosette forming a rectangle. I believe it is the frame from the 120 sec set of images.

I cannot ascertain the cause of this nor how to mitigate. I have tried a few post processing techniques, albeit not thorough as I am still a newbie.

Attached a jpeg....I hope it shows. Any advise?

BlockHead

There appear to be someissues here. First there looks like there is walking noise. Are you dithering your data? second, I am surprised at the level of noise. How many images did you acquire in each set of data? (I would guess less than 20.) Next, the difference in SN between the frames shows between the two because your two sets are rotated with respect to one another. With enough frames and proper Local Normalization I would not expect to see this. Finally, based on the noise and lack of normalization, I am not going to rule out that some (all) of the data is not calibrated properly.

The normalization is the thing I really don't like about this. I assume you let it automatically pick the images used for the reference- but you need to be certain it is picking images from only set set (60 or 120) data and it is likely the 120sec exposures would be best. To be certain, using the interactive mode in this case can help. 

This is an advanced topic since you are mixing data sets that are not of the same quality and different in other ways. I think this is a bad trap for a "newbie." Part of the issue is that the acquisition decisions are affecting your processing strongly. You usually want to avoid this.

-the Blockhead

Professor

There are 140 60 sec light frames and 47 120 sec light frames after blinking.  There were 16 60 sec frames rejected in WBPP process "Bad Frames Rejection" process.

The images were dithered.

I did let WBPP pick the images. I will need to study the normalization process and master this problem. Combining multiple session images will become more of a routine.

I had forgotten to mention, I did run the light frames in WBPP with no calibration frames.  i still got the same number of rejection frames and the final integrated image still had the 45 deg angled frame lines running thru the image, but not as evident. It was quite hard to see them, but they were there. The calibration frames made them much more evident.

Thank you.

BlockHead

Yes, you need to calibrate the data! 

It appears to me you showed me integrated non-calibrated data. 

This is *not* the AdamBlockStudios way. LOL

So, if you calibrate the data properly the walking noise will likely be less of an issue.

The other issue (which is a complication of the rotated frames) is the LN normalization.

If you do not rotate your frames... this also will not be an issue and things will be easier. 

So at the moment we are talking about proper calibration and optimal acquisition.

That being said, you can accommodate your data that you have in hand. 

I will shortly be releasing a video that demonstrates interactive LN. It is already a part of the WBPP series as well as in Fundamentals as a stand alone section. 

-the Blockhead

Professor

The data of the jpeg image I posted is of calibrated data. Running a non-calibrated session was stricly an experiment to see if the condition appeared which it did.