Regional HFR

The whole-image HFR is a great feature, but I suggest the user should be able to select only a region of the image for calculation. HFR now includes galaxy cores, saturated stars, and bright extended objects, which throws the image statistics off and can cause autofocus inaccuracy or failures. Selecting a region for HFR would result in faster calculation and better focusing by using only known unsaturated stars.

1 Like

Just a thought, if the focus image was plate solved then it would be possible to identify the blobs that are stars and eliminate the ones that are extended objects. Then the real stars can be used for focusing.

The obvious problem with this is that if you aren’t focused it may be difficult to plate solve, still there may be something that can be used.

Chris

@Chris – That might eliminate HFR on the extended objects, with the caveats you mention, but HFR includes saturated stars in its calculation too. Those are not going to give realistic HFR values or probably respond properly to autofocus. Most astrophotos will contain a dozen or two saturated stars.

I think letting the user select an area they know won’t include saturated stars or extended objects would be good. It shouldn’t be that hard to do, either.

How about using outlier rejection?
Like if a star is oversaturated and doesn’t give a good V curve comparable to most stars in the image it will be excluded.
It will of course not be possible for the first images, but it could be applied when calculating the focus position before the final move back to focus.

@Xplode – I bet they don’t calculate a V curve for each star during autofocus. They probably use the aggregate. Keeping track of the V for each star would require hundreds of times more memory and computation, but maybe that’s not out of the question.

Probably not, but it could definitely be an option to do for computers with the resources to handle it.
It should be an optional thing so slower computers won’t get bogged down.