Improve optimal SNR calculation when signal is injected at the start of segments and hence cut off

The current calculation of optimal SNR for injections is imperfect when the signal is injected at the start of segments and cut off. As a consequence the injection would appear in the list of high SNR missed injections follow-up page, but I think ideally it should avoided by getting the actual in band optimal SNR 