Fix the matching method of the L1 and offline muon in the TrigMuonMonitoring
I want to merge this change into 22.0.
In the TrigMuonMonitoring, when calculating the L1 efficiency, the threshold number was used to determine if the offline muon was passing L1. However, this is not a good idea. In this MR, the judgment method of L1 pass was changed to whether initial RoI was created by L2SA step. The initial RoI is made only when passed L1, so it is a more correct judgment method.
This change slightly changes L1 efficiency and L2SA efficiency wrt upstream. https://mwakida.web.cern.ch/mwakida/ATR-25631/test1/MuonMon/Efficiency/ZTP/HLT_mu26_ivarmedium_L1MU14FCH/index.html