NaIγ Fingerprint Correction And Its Artificial Neural Network Recognition

γ fingerprint identification technology is one of the main methods to identify nuclear materials,and in the radioactive source security management,nuclear emergency monitoring,nuclear terrorism prevention and nuclear weapons verification in nuclear disarmament and other fields have been widely used.γ fingerprint acquisition system mainly includes HPGEγ spectrometer and NaIγ spectrometer.Compared with HPGE γ spectrometer,NaIγspectrometer has the advantages of high detection efficiency,easy maintenance and low cost.However,the temperature stability of NaIγ spectrometer system is poor,which restricts the application and promotion of NaIγ fingerprint identification technology.Therefore,the effective elimination of NaIγ fingerprint line drift,to expand the application of NaIγfingerprint identification technology has important practical significance.In order to solve the problem that the spectrum drift of NaIγ spectrum is serious and it is difficult to meet the application requirements of γ fingerprint identification technology,this thesis proposes a method to correct the NaIγ fingerprint drift based on the theory of stochastic signal system transformation.The theoretical deposition of the measured γfingerprints in the detector was performed by using the inverse transformation of the system.The spectral drift,deformation and distortion caused by the system parameters and instability of the γ spectrometer were eliminated.At the same time,the artificial neural network method was used to identify the simulated and measured nuclear material γ fingerprint before and after the correction,and the feasibility and effectiveness of the correction method were verified.Specific work contents are as follows:(1)The feasibility simulation of the system transformation correction method of NaIγfingerprint.Nuclear material NaIγ fingerprint was simulated by Monte Carlo simulation technique.One group is the drift γ fingerprints of the same known nuclear material under different system parameters,the other group is the non-drift γ fingerprints of the same nuclear material with little difference under the same system parameters.The proposed correction method was used to correct the simulated γ fingerprints,and the γ fingerprints before and after the correction were identified to explore the feasibility of the correction method.(2)Experimental verification of the validity of NAI γ fingerprint correction method.The NaIγ fingerprint correction method of the effectiveness of the experiment on the basis of the work(1),the standard source in laboratory as the research object,using the γNaIγ spectrometer to measure the drift fingerprints and little difference between nuclear material γ fingerprints: γ fingerprint is obtained by adjusting spectrometer system gain drift,by changing the measuring time there’s little difference between nuclear material γfingerprints.The proposed correction method was used to correct the measured γfingerprints,and the γ fingerprints before and after the correction were identified to verify the validity of the correction method.The results show that the γ fingerprint correction method based on the stochastic signal system transformation theory can effectively eliminate the peak shift and longitudinal spectral intensity change of the NaIγ fingerprint,reduce the identification lower limit of nuclear material NaIγ fingerprint,improve the recognition confidence,which can provide theoretical and technical support for expanding the application of NaIγ fingerprint identification technology.