scale hyperspectral datasets

[borondics]

Currently the spectral reader simply imports the row and col numbers.

However, it would be even nicer if we calculate the real (relative) positions and add two extra columns into the dataframe that would be based on the number of steps in X and Y directions combined with the information from the metadata 'ScanArea'.

Also, make sure that the dataset center is shifted with 'Scanner Center Position'

[ngergihun]

Sure, good idea, we already have this in the orange-snom reader, but then a calculation can already be done here.