Why was this using a divide to calculate the difference?

Lack of clamping while reading a pixel from a palette resulted in reading invalid memory

Initial CMake support

Checking if the type string contains paletted instead of just the current image
leading to palettes attempting to be applied to images without a color buffer

Creating a paletted image along with other images broke it

I will tidy this up in the next few commits

Processing more than one image at once didn't work (quite bodgey)
MCIR would run even without a map
Error handling for map/raw images that don't exist

Allows custom filenames for the output image
Fix formatting in help text

The getsamples function is now able to only get samples from only one channel

Also improve land color in MCIR images

Lines are now only skipped before there is a detectable sync marker, but not after.
Simply removing the part responsible for recalclating pixel offset

Add linear equalise
Add image flip for southbound passes
Completly remove old falseclor in favour of a much simplier method

Add missing newline to usage text.

Change calibration to better suit constant gain
Add -Wextra to Makefile
Add link to model of the PLL
Max number of rows for the image is now increasible

Also change palette to something nicer

Add a denoise filter
Fix temperature calibration
Tidy up some files
Improve MCIR

Add other image modes back into ImageOut

Add the ability to open a map overlay generated by wxmap (still unstable)
Remove the ability to turn off calibration
Add MCIR images
From now on large changes to neaten up the codebase will be taking place

Since distribution maps now use PNG

Add the ability to open 8bit PNG's instead of audio files, this is useful for things like playing around with false color paramaters due to reduced processing time
Also reduce some buffer sizes to more sane values

Also made some floating-point math more explicit