making normalized correlation faster

[hornig]

https://github.com/aerospaceresearch/PlaneSpotting/blob/5c44881bde6b35c86bfe6e27673ae064346fc8a4/main.py#L5

making normalized correlation faster.
numpy has no normalized correlation, unfortunately. so I built our own version from as many numpy functions as possible.

there should be a better way. we won't be the first one having this task. find a better and thus faster way to find the preamble starts.