Investigate event normalisation (rather than volume) in IntegratePEaksMD

[RichardWaiteSTFC]

IntegratePEaksMD subtracts background counts from peak intensity by normalising by the ratio of the volume of the peak and background regions. However, the binning is not uniform in Q-space, so it would perhaps be better to normalise by the number of bins/events (as is suggested in FindPeaksMD