Development of pileup recovery algorithms by peak detection method of digital gamma ray spectroscopy

The Pulse pile-up in gamma ray spectroscopy is a severe problem because it degrades both energy resolution and detection efficiency, especially at high count rates. This paper presents two algorithms for signal recovery of peak pile up. These algorithms depend on experimental acquisition of gamma radiation signals with analog to digital converter at sampling rate 16MS/s. The first algorithm depends on a peak detection method that determines the peaks' parameters, which are recorded into a two rows matrix representing the amplitude and position of the peaks. The second algorithm depends on a peak sensitivity method, which determines the location of the maximum peaks using a local extreme method. The variance between each two adjacent points are calculated to find indices of nonzero elements that are larger than zero to determine the maximum amplitude of each peak. The resultant peaks correspond to peak location. These two methods are evaluated by comparing not only their parameters but also their fitting error. Moreover, the influence of these algorithms on the obtained spectrums are analyzed and evaluated. The results show that the second algorithm outperforms the performance of the first one due to the efficient recovery of overlapping peaks.