Sequence Recover after PHD2 loses star

Tal Akerman

This is a major enhacment and will greatly improve shooting with automation, especially on partly cloudy nights.

Would love to see it happens.

2020-09-26T14:47:56+00:00

Tal Achituv

Agreed. Would save so many lost images.

Needs to take into account clouds. So repeat platesolve until successful. would be nice to verify number of stars before platesolve in such instances to avoid hammering the servers with cloud images.

‌

2020-09-27T13:36:26+00:00

Ziv Arazi

That will be a very positive influence of my work flow after a star lost on phd

2020-09-27T14:07:04+00:00

Tal Achituv

Some discussions offline made me think this might need more details, as some people confuse this with the ‘recover lost star’ feature.

When PHD2 loses a star, NINA already can recover (if configured to do so) by automatically re-selecting a star.

However - when the guide star has been lost for long enough for the whole frame to move significantly (which usually isn’t very long, especially since PHD2 sometimes tries all sorts of aggressive moves due to noise/clouds), then a whole new platesolve-slew loop is required to get the frame back. Only once the framing is correct, should PHD2 resume guiding.

It would perhaps be wise to implement a frame-comparison feature to help figure out the offset between frames, and unless subsequent frames can be shown to be within a few pixels from each other, then there are either still clouds or PHD2 guided us away.

If we can distinguish between clouds and no-clouds (such as by star count, skyglow, or perhaps some cloud-specific algorithms) then we can implement auto-reframe once clouds are gone.

some of this might be better implemented in PHD2, since cloud detection can work much better when we have ~1S exposures rather than long exposures. (motion vectors, brightness changes, etc.)

‌