I tried to stitch the Everest horizon b/w full res. pan from L7 pancam pics...



(4,08 MB) http://www.greuti.ch/spirit/Spirit_pancam_sol620-22.jpg

The pics were first processed with Michael's anti-vignetting programm, stitched with PTgui and finally processed with PS6.

But now I'm "deadlocked" when adjust the frames together.
The problem is due to the brightness gradient that continues over the whole pan from left to right, from bright to dark.

Has anybody an idea how one could overcome with it? A solution would be if one could correct this brightness gradient, likewise such a simulated brightness gradient in the sky that we sometimes use in pans - only contrary corrected to a uniform gradient.

Tman,

I don't have the answer to you, but can say that autostitch doesn't seem to suffer this problem.

Here is a first trial with Everest pano (66 of 81 images) -- absolutely nasty, off course -- in which we can observe the obvious vignetting effect but not the gradient visible in your PTGui stitch.

Click to view attachment

It seems that its gain compensation mechanism solves this issue, but then the solution to your problem would be first of all to know the autostitch algorithm and then to implement it. Anybody knows it?

Could be Tes, but in such automatically adjusted pans you mostly get an "dense" horizon, don't you, because of the high contrasted frames. I could it get too when processing with PhotoShop - but if it works, I would prefer avoid such damages of details.

could it be that in this case, the brightness variations are not only due to Vignetting
(within one frame) but also somehow different exposure conditions accross individual frames ...

(I wonder if the well known auto-contrast stretching "feature" of the published MER JPEGs also applies to individual frames of a panorama series ??)

So the problem for the global gain compensation would be to map all the brightness variations of all frames into
the dynamic range of a single image (most common problem would be a loss of
subtle brightness variations in the sky ....
So no easy solution for this one except to carfully do the brightness adaption "by hand" for the individual frames ...

So there are a couple of possibilities but a lot depends on how the image exposure is handled. Do we know how the exposure times are worked out? Is there an auto-exposure feature within the cameras or is the exposure fixed on a daily basis following some earth side calculations using predicted insolation and recent tau measurements?
It wouldn't be trivial but it should be possible to work out a method to correct the exposure on the images for changes in insolation over the period of time that the panorama images were taken.

Hey, fortunately I found a useful solution for my problem in PhotoShop itself that works not bad.

http://www.unmannedspaceflight.com/index.p...indpost&p=22745

You have to copy such a layer part of the pan and then in this layer to create a b/w gradient that lies as transparent and adjusted mask over the original part of the pan.

Due to the different exposure of some pics you can correct these differences more or less in "picture calculation" (Bildberechnung) in PS. That problem I've had was more caused by the direction of the sun during the foto sessions I guess. Now this method with such a mask should works for each several frame too.