First attempt at image comparison using simple stats

[kcartier-wri]

Crude prototype of using simple statistics to compare rasters with different resolutions.

[kcartier-wri]

Here's the prototype of "fun-with-statistics" version of the raster comparison function. I may be able to make it more justifiable and robust with further refinement.

One concern is that the tolerances must become more extreme as the raster coarseness increases and value variability decreases. Another is that the code only does a set-based statistical analysis and completely ignores the spatial aspect of the data. Maybe we could identify some maxima/minima and check spatial (or grid) proximity.