psImageRebin - clarification of functionality.

[robert.desonia@…]

I need clarification on the expected functionality of psImageRebin. Is the
scale assumed to be 1/n, where n is an integer? If not, the issue of how to
handle subpixels comes into play and makes the routine an order of magnitude
more complicated to implement (area weighted statistics?).

If scale is to be 1/n, I suggest defining the scale parameter as the integer n
above, i.e., the number of pixels to bin together in each dimension.

-rdd

[robert.desonia@…]

Eric pointed me to the ADD; Does ADD section 1.2.1.1 apply to how psImageRebin
should function as well?

[eugene]

I agree: this function is doing integer rebinning, not arbitrary remapping (we
are defining that one for the future for other purposes).

The scale should be the binning factor, and it should be an integer.

(there should also be the opposite function, psImageResample, which expands each
pixel to NxN pixels. This would require some interpolation options.)