What's going on here? Convolution, STF, then undo convolution

[carl]

Hi folks,

I'm doing a particular project using SDSS data trying to detect faint tidal streams. I've discovered that the following process shows them up, but I don't understand what's actually happening in PI terms.

1. Convolution (gaussian, 7SD)
2. STF AutoStretch (Boosted)
3. Undo convolution

Ideally, I'd like to automate this entire process so is it possible to "undo" a convolution in a process container or JavaScript or something?

Secondly, can someone explain how the undo works after the STF has been run?

Thanks,
C.

[RickS]

All you are doing is applying a stronger screen stretch to the data.  STF calculates the autostretch on the basis of the smoothed data then you undo the convolution and keep the stretch.  The screen stretch doesn't get undone because it didn't actually modify the image data - only the parameters used to display it.

[carl]

Thanks Rick. I thought it might be something like that.

Is there a way to script the "undo convolution"? I would like to automate this process over a few hundred images.

[RickS]

Hi Carl,

I expect it is possible to do this in Javascript but you can probably get a similar effect by adjusting the STF autostretch parameters for a more aggressive stretch without the need to convolve/undo.  Ctrl-left click on the STF "radioactive" icon to find the autostretch parameters.

Cheers,
Rick.