The Overlapping Peaks Decomposition Method of Nuclear Spectrum Based on the Statistical Clustering Model

The full energy peaks always overlap each other when carrying out radioactive spectrum measurement. In this paper, combining GMM statistical clustering model with EM iterative algorithm, the overlapping peaks of nuclear spectrum with no background has been decomposed. No matter how many overlapping peaks we can use this decomposition method, and ensure that the statistical errors close to zero. It has great practical value for radionuclide with a quantitative or qualitative analysis.