CS4: Combining LR images to increase resolution?

Question

I've got a number of images of the same object, all at the edge of frame in a photo of something else! There is some text on the object that I want to read.
Is it possible to combine these low-resolution images to increase the resolution? Like Merge to HDR, but to HR?

Answer 1

It's a concept astronomers use.  Look up the term "Drizzling", for example, as related to the Hubble Space Telescope.
Assuming your subject hasn't moved, you create hyper-accurate data by overlaying multiple exposures, then deconvolve (sharpen) it aggressively.  A bit like having a camera from 10 years from now. 
Try this:
1.  Upsample each of the images to, say, 200% of original size.  I might even suggest 400% if you have the computer resources.  You can, of course, crop all but the part you want to enhance.
2.  Layer them one on another in a single document.
3.  Use the alignment feature (Edit - Auto Align Layers).
4.  Set the transparency of each layer to 1 over the layer number, with the Background being layer 1, the first layer above that being 2, and so on.
5.  Once you have all the images precisely aligned and the proper transparency, flatten the image.
You'll now have a very accurate single image of your subject.  There are commercially available deconvolution programs (check out Focus Magic for example), and you can experiment with various sharpening techniques and programs.  All should do well, since (if everything came together well above) the data will have less noise and better accuracy than any single image.
Sorry to say it's not simple - i.e., there's not a single button to press like Merge to HDR - but it should be doable.
-Noel