Warped Image --> Warped Intensity?
MartyMcFly
Sat, 07/13/2013 - 02:19 am
doing photoemission, I get angle-resolved data on a CCD detector. These are my images. The measured raw images has afterwards to be warped for some reasons. This warping can be done easily by IGOR's ImageInterpolate routine.
Now the question: Is it, in a simple way, possible to transform also the intensity (i.e. the pixel values in z direction) according to the transformed area of such image? That means: if area A is warped to a new area B, the pixel value average P_A in the first area should be the same for that of the transformed area (P_B).
Physically, it would mean that the numbers of electrons impinging the CCD detector (i.e. the intensity) should not change for the warped image.
Would be nice if anybody has an idea!
Marty
Interpolation will conserve the pixel value average. To conserve the pixel value integrals instead, you need to multiply the interpolated pixel values by the area in the old image covered by the new pixel. The change in pixel area is approximated by the divergence (in pixel units) of the displacement field used to warp the image. I am unsure of the actual displacement fields used in Igor's ImageInterpolate Warp method (I don't have a copy of that ILM Technical Memo #1030), but if you make two copies of your image, set one of them equal to p and the other equal to q, and morph each of them identically to the original image, that will give you the displacement fields you need.
I hope this is enough to get you started.
-- Sven
July 13, 2013 at 06:17 am - Permalink