Fingerprint Identification Of γ Radiation Spectrum Based On Artificial Neural Network

Gamma ray spectrum analysis technology is widely used in environmental sample survey, geological prospecting, radioactive waste monitoring. At present, nuclear arms control and disarmament has become a worldwide trend. Nuclear disarmament means all nuclear-weapon states must decrease nuclear weapons in number, using agreement or convention to provide for the number of nuclear warhead possessing by all the countries. The main method of nuclear disarmament is reducing nuclear weapons by a big margin through a verifiable and irreversible form. In order to make the destruction of a nuclear warhead is the specified nuclear warheads, we need to distinguish the nuclear warhead through the radiant fingerprint. So the technology of y energy spectrum fingerprint played a key role in identified the nuclear warhead that had been reduced.y-ray is a radiation form of nucleus energy level jump. It is the nucleus stimulated to produce ray radiation, when the nucleus α-decay and β-decay. There are many y-ray detectors, but I used is HPGe detectors in the experiment. y-ray can occur three kinds of effects in the detector, that is photoemission, Compton effect and pair effect. Accordingly, y energy spectrum consists of full-energy peak, Compton plateau, single(double) escape peak, noise background and so on. Each unique radioactive nuclide corresponds to the y energy spectrum, that is, different radioactive nuclide has different characteristics. We can consider the different characteristics as different fingerprint, so the y energy spectrum can be defined as "y energy spectrum Fingerprint".The artificial neural network, is a technique of biological neural network simulation. It is based on connection of artificial neuron. It is an intelligent system which can give suitable analysis of the input signal processing. The artificial neural network has the characteristics of distributed parallel processing, nonlinearity, robustness and fault-tolerance. So it is widely used in fuzzy identification, intelligent control and nonlinear signal processing. BP network and RBF network are two commonly used artificial neural network, BP network is using error back propagation error back propagation algorithm of the learning process, global approximation training process, And RBF network use local approximation, and parameters were needed is small, So the RBF network prediction accuracy is far greater than BF network. In this paper, we use BP neural network and RBF network to gamma energy spectrum identification for different age groups and different size nuclear warheads, Because of the big parameters of the BP network, and slow training speed, low accuracy, So it can’t accurately identify for different gamma ray spectrum,and the small parameters of the RBF networks, training speed, high precision, so it can be well used for identification about different gamma ray spectrum.