Wavelet Denoising of Mobile Radiation Data

The investigation of wavelet analysis techniques as a means of filtering the gross-count signal obtained from radiation detectors has shown promise. These signals are contaminated with high frequency statistical noise and significantly varying background radiation levels. Wavelet transforms allow a signal to be split into its constituent frequency components without losing relative timing information. Initial simulations and an injection study have been performed. Additionally, acquisition and analysis software has been written which allowed the technique to be evaluated in real-time under more realistic operating conditions. The technique performed well when compared to more traditional triggering techniques with its performance primarily limited by false alarms due to prominent features in the signal. An initial investigation into the potential rejection and classification of these false alarms has also shown promise