SubframeSelector Feature Request

[fulatoro]

Good Day,

Great job on SubframeSelector/FWHMEccentricity.  I have been using it trying to resolve image quality issues that I suspect are related to tilt.  I can see based on the Eccentricity maps that I have issues to resolve...What would be really nice would be to have an eccentricity map resized to be the same size as the light frame so that I can overlay the eccentricity map over my light frame to more directly correlate the eccentricity map with the actual stars in the frame.

I realize that this a bit more complex since the eccentricity map would have to be resizeable to match the current light frame window size.

Anyhow, just wondering if this would be worthwhile addition.

Thank you.

[mschuster]

Thank you. As a workaround, please try scaling the map up to matching size with the Crop and Rescale processes. The result may be slightly blurry, but should remain readable well enough. The loss of accuracy compared to a true hi-res map will be negligible IMO, due the inherent area averaging in the star measuring and map mapping processes.
Mike

[fulatoro]

Makes sense. I will try that. Since I am working with DSLR images, I my workflow has been to first debater the raw light frames, then extract a luminance which I the feed into the script. 

Is this the recommended way of doing it or should I start from the calibrated frames?

I am still learning PI so I have been thinking of a way to automate the process, one option would be to use process containers, the other one once I get a better handle on the API would be to add an input parameter to SubftameSelector to specify that the frames are cfa and modify the code such that it debayers the image and extracts the luminance component.

Is this viable?

Thank you.

[mschuster]

On my monochrome sensor frames the calibrated and uncalibrated measurements are similar for the same filter, IMO either way is OK.

Using debayered luminance sounds fine.

I don't recommend adding debayering to the script. IMO better to keep the processes separate, not duplicate functions, and avoid testing and maintenance work.

Mike

[whwang]

Hi,

Please allow me to add another feature request to the FWHMEccentricity script.
In the current version, the "support" button generate FWHM and Eccentricity maps.
I think it will be very useful to also generate a vector map that shows the direction
of the elongation of stars.  This should provide a even more powerful diagnosis
about the problems in imaging.  For example:

1.  If the elongation is caused by tracking problem, the eccentricity and direction
of elongation should both be uniform across the image. 

2. If the elongation is caused by coma, field curvature, and astigmatism etc, the
direction should be either tangential or radial.  The pattern should be symmetric
around the center of the image. 

3. If the elongation is caused by field rotation (guided exposure with poorly
aligned equatorial pole), then it should be radial but the center may not be
in the center of image.

4. If the elongation is caused by tilt of sensor plane, the eccentricity should show
a uniform gradient, but the direction should be constant.

5. If the elongation is caused by collimation error, then anything can happen.

Cheers,
Wei-Hao

[mschuster]

Hi Wei-Hao,

I agree a direction map would be useful. An early version of the script actually did generate such a map, but the technique I used was a bit of a hack and the results were not always reliable on my under sampled frames, so I removed it. Maybe I will try again sometime.

Mike