SP3 Star detector

[Forum_2017]

Hi Fabio,

I think the 'star detection' parameters in the 'preferences' do not seem to be applied immediately (you have to reload AA6). This may be OK as normally the preferences are only set once and left. I just happened to be trying to tune 'star detection'.

Another issue with 'star detection' is that some star that are not detected but are within the preferences range. I think there might be a 'star profile' issue. I can see from the 3d view, of the non detected stars, there are possibly 2 peaks on the same star image. However, it can be selected manually and when selected it is within the star detection parameters.

Clearly star images can have more than one peak if the atmosphere is moving.

I think doing any form of smoothing of the star image could have an impact of star ADU and thus accuracy of measurement.

All the best

John

[Forum_2017]

Hi John,
nothing have changed between SP3 and SP2 in star detection. Stars with noise and multiple peaks may not be detected by that script command, although they are detected by "Find stars" and the other commands of Astroart. Only a patched version of that script command could solve the problem.
Fabio.

[Forum_2017]

Fabio,
I am sure nothing has changed in star detection SP2 to SP3. I have recently been trying to improve the accuracy and reliability of measurement of my system.
When doing 100 measurements of UU AQR using 3 ref and 1 check star as the star dimmed and sometimes when atmosphere made star shimmer by 1 or 2 arc sec it would fail to find the star. Which left some critical gaps in the time series.
Is there any way to set the ref flag and value from a script? Or to use the same star detection as Image.Stars.Find(x),
Then maybe I could set the preferences to eliminate non stars or very dim stars, Find(10000) and then search the list (normally about 100) for the correct RA DEC within say 2". Set ref and value and do Image.Photometry. Just for those that fail the original method.

When manually selecting the failed stars and inserting the necessary values the photometry comes out very close (better then 0.05mag)
The alternative is to bin that image 2X2 and re apply the photometry. However, I have noticed that binning does reduce the accuracy and improve detection. I am sure this is due to smoothing the star profile.

Any ideas?

Regards,
John

[Forum_2017]

Hi John, it's faster for me to patch that script command so that it finds the same stars of the other commands of AA. Indeed the problem was stars with two peaks. The correction is ready yet, it will be included in the next release. For a prerelease just write me by email.

[Forum_2017]

Ok the answer is to Bin2X2 as the error only doubles but the point is still reasonable.
Example:
JulianDate Filter VarAbsMag VarAbsErr ExpLen CmpStar RefMag CmpStar RefMag CmpStar RefMag
2458056.327578 V 13.6350 0.0170 100.0 121 12.084 132 13.167 129 12.865
2458056.328919 V 13.4650 0.0143 100.0 121 12.084 132 13.167 129 12.865
2458056.330264 V 13.5620 0.0161 100.0 121 12.084 132 13.167 129 12.865
2458056.331605 V 13.5250 0.0226 100.0 121 12.084 132 13.167 129 12.865 *
2458056.332948 V 13.5130 0.0220 100.0 121 12.084 132 13.167 129 12.865 *
2458056.334292 V 13.4800 0.0153 100.0 121 12.084 132 13.167 129 12.865
2458056.335635 V 13.4930 0.0213 100.0 121 12.084 132 13.167 129 12.865 *
2458056.336977 V 13.4660 0.0143 100.0 121 12.084 132 13.167 129 12.865
2458056.338325 V 13.5220 0.0148 100.0 121 12.084 132 13.167 129 12.865
2458056.339666 V 13.5770 0.0236 100.0 121 12.084 132 13.167 129 12.865 *
2458056.341010 V 13.5590 0.0234 100.0 121 12.084 132 13.167 129 12.865 *
2458056.342354 V 13.4830 0.0147 100.0 121 12.084 132 13.167 129 12.865
2458056.343700 V 13.5220 0.0152 100.0 121 12.084 132 13.167 129 12.865
2458056.345041 V 13.7320 0.0272 100.0 121 12.084 132 13.167 129 12.865
2458056.348115 V 14.3460 0.0313 100.0 121 12.084 132 13.167 129 12.865
2458056.349459 V 14.4230 0.0327 100.0 121 12.084 132 13.167 129 12.865
2458056.350802 V 14.8720 0.0497 100.0 121 12.084 132 13.167 129 12.865
2458056.352147 V 14.8170 0.0450 100.0 121 12.084 132 13.167 129 12.865
2458056.353488 V 14.6060 0.0382 100.0 121 12.084 132 13.167 129 12.865
2458056.354829 V 14.2550 0.0285 100.0 121 12.084 132 13.167 129 12.865

The ones with asterisk are failures bin2x2

So I think the error increase is within my acceptable level.

Regards,
John