Key Technology Research Of γ Spectrometer Aimed At Multi-rotor Type Aircraft

As military nuclear weapons, nuclear materials and technology diffusion is becoming more and more serious, and the typicalsuddenof the radiation accident and radiation terrorist attacks,they are great threat to social security and national politics and economic, Multi-rotor aircraft with nuclear detection equipment in the application of nuclear environment had gradually highlighted its advantages and immeasurable effect. This paper focused on the problems of radiation detection, recognition and treatment, carried out a series ofresearch work aboutkey technology of multi-rotor aircraft with gamma spectrometer whichclosely combines the theory and practice. Main content includes the following aspects:On the foundation of the research about theories and resultsof the digital spectrometer, we can concludethe main factors affecting the performance of the gamma spectrometer of the multi-rotorand the key technical problems need to be solved.For multiple-rotor aircraft platform, compare and analysis the traditional nuclear spectrometer digital processing method, choose synthesis method, time comparison method, and average method as the platform of gamma spectrometer digital processing method, lay the foundation for the analysis of energy spectrum.Focused on thedefectthat traditional line smoothing method will cause distortion of spectrumwhenthe parameter of the filter is improper or the frequency ofsmoothing is too much, this paper puts forward a recursive least squares method of spectral smoothing. Take advantage of the principle of adaptive filter, using the fast kalman recursive least squares method, and through the forward and backward predictor update the filter coefficient, realizing the processing of energy spectrum data smoothing.Through the experiment comparedthis method with polynomial least squares method and least mean square algorithm in qualitative analysis and the results demonstrate that this method can reduce the noise in the spectrum, and can keep the characteristics of spectral.In view of the problem that traditional neural network training effect in the nuclide identification is weak, easy to fall into local minimum, slowly convergence speed, this paper nuclide put forwardanidentification method based on probabilistic neural network.Thefeatureof energy spectrum data smoothed will be extractedby Hilbert--huang transformation, then the peak width, characteristics of the energy intensity and peak area characteristics of information as a probabilistic neural network input, then set up training and testing samples for nuclide identification.The experimental results demonstrate that this algorithm can deal with thesituation of a single nuclides, mix nuclide and unknown species, can be applied to safety monitoring and uncontrolled radiation detection, and other areas of the rapid nuclide identification.Finally, set up atest platform for key technologyof multiple-rotor aircraft gamma spectrometer system, this platform has a preamplifier, adjustable gain amplification and the function such as data acquisition, and it can verifythe feasibility and the rationality of the optimal digital nuclear signal processing scheme which come from the theory.In case of the nuclear accident emergency, the use of multiple-rotor aircraft with nuclear detection lay a solid foundation for nuclear accident scene investigation and barriers to exclude.