There seems to be some problems with the the maths and the physics here.

If you go back to the basics, the WPI has a 2.04um pixel, whilst the K20D has a 5.03um pixel size.

Furthermore, if you compare the crop factor for the WPI vs the APS-C camera, then the WPI has a crop factor of 4.5 x 6 compared to the APS-C value of 1.54 x 1.54.

So to have a valid comparison in image size and then compare pixels the 18mm focal length on the K20D is 27mm vs 81mm on WPI.

Me thinks it is not a fair comparison. You should therefore have the WPI set to a focal length of 6mm to have the same field of view on both cameras.

Now do your comparison.

The second issue to question in your argument is the ability of the pixels to resolve the photons striking them. Smaller pixels will have poorer signal to noise ratios. A smaller pixel will have less photons in it and the signal to noise (the square root of the number of pixels striking it) will therefore be signficantly higher.

Therefore, the WPI, with the smaller photoelement will have MORE noise as it has a physically smaller dimension to capture those photons.

In practice, this means that for smaller sensors you need far better lenses. Is the WPI lens better than your SLR lenses?

There will also be an effect on contrast and spatial resolution.

If you want a proper analysis of this check out the following links:

DSLR SENSOR SIZE AND PIXEL DENSITY Clarkvision: Does Pixel Size Matter http://isl.stanford.edu/~abbas/group/papers_and_pub/pixelsize_talk.pdf
So please, go back and take your images again - then we can validly compare them.

BTW, if you've got access to a K100D then the comparison there would be valuable.