25/08/2022 - 08/09/2022 Meeting Notes

[Victoria-Samboco]

25/09/2022

In the meeting of 25/08/2022 I presented the results I got from 1GC and SELFCAL step with caracal. But the images was showing showing part of a very bright source where was concluded that the following parameters must be changed to improve the image:

1. image size from 6000 to 10240 to include the bright font that was halfway up in the image and therefore might not be being cleaned up.
2. nr of channels from 5 to 8;
3. automask from 20 to 10 to improve the cleanliness of the image and with the auto-mask of 20 we had a lot of artefacts;
4. spectral-fit from 2 to 3;
5. img_specfit_nrcoeff: from 3 to 2.

After modifying the above parameters I ran SELFCAL again, but it was verified in the result that in the image generated by the CORRECTED_DATA column a second source appears that was considered a fake source and the model also had this source and black dots (negative regions on the sources) (already discussed in Issue #8 ).
With that it was concluded during the discussions in issue #8 . That I should use different deep masks generated from breizorro to analyse if the artefacts and the fake source keep appearing in the model.

08/09/2022
In today's meeting I presented the results of the images generated from different masks (mask2, mask5, mas10, mask15 and mask20) as well as the models. Where it was seen that with the result of the fake source brezoirro masks were not in the model which is good. But it was also seen that the model with mask2 still contained a lot of artefacts (because it was very low) the model with mask5 still contained some artefacts and mostly (almost all the amount of artefacts was not present in the model).
With that, it was suggested for the follow-up of the work that:

1. Generate a mask with threshold 7;
2. Run SELFCAL step again with the new generated mask (with threshold 7) and that should be okay.

After that I can proceed with the Sun imaging.

[o-smirnov]

@Kincaidr's efforts made me realize the obvious, @sihle20's transient search needs a pretty much identical pipeline to a point (i.e. best possible selfcal, all sources subtracted). So @Victoria-Samboco @sihle20 please work together on this.

[Victoria-Samboco]

Okay, we will do it.

…

On Fri, 9 Sep 2022, 10:16 Oleg Smirnov, ***@***.***> wrote: @Kincaidr <https://github.com/Kincaidr>'s efforts made me realize the obvious, @sihle20 <https://github.com/sihle20>'s transient search needs a pretty much identical pipeline to a point (i.e. best possible selfcal, all sources subtracted). So @Victoria-Samboco <https://github.com/Victoria-Samboco> @sihle20 <https://github.com/sihle20> please work together on this. — Reply to this email directly, view it on GitHub <#9 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWYM2BNU2V35QV77AVZECC3V5LW53ANCNFSM6AAAAAAQH4RF4M> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[Victoria-Samboco]

Greetings

I had as a result of SELFCAL using threshold 7 with the mask created with breizorro these two images. The first image is the restored image and the second is the model (both from the CORRECTED_DATA column). As desired the fake source and most of the artefacts are no longer in the model but we still have black dots on the sources. I'm not sure if I understand why they appear and if they are a problem or not.