I am a geo guy and exploring using processing to make some of my output a little more pretty.  I am currently exporting a flat map with a 3d perspective as an image and bringing it in to processing.  The problem I am having is trying to convert the pixels to geo when it is not a 2d rectangle, almost like this twitter map from visualcomplexity:

 

 

Most of the methods for converting pixels to geo I have seen, use bounding lines (top, bottom, etc.).  Is there a way to define them using bouncing points (upper left, upper right, etc.) and then use those to do the geo conversions?  Or is there a better way to do that that I am not aware/thinking of?  I am new to programming, but more familiar with db style language, so I figured I come to you guys for help.  Thanks!