论文部分内容阅读
目前的民航安检X射线透视设备无法直接检出塑料炸药。国际上初步研究有效的γ射线共振技术可以透射瞬时测定行李中的物件的氮、氢、氧、碳含量。为与此技术配套,本工作应用对小样本集统计预报特别有效的支持向量机(support vectormachine,简称SVM)算法根据样品的氮、氢、氧、碳含量判别常见民用品和炸药,并用留一法比较SVM,Fisher法和人工神经网络算法的预报效果。结果表明SVM算法误报最少,且对所列炸药无一漏报。据此建立了炸药判别系统软件的原型,在实验室中模拟测试结果良好。
The current civil aviation security X-ray fluoroscopy equipment can not directly detect plastic explosives. Internationally preliminary research effective γ-ray resonance technology can transmit instantaneous determination of the contents of luggage nitrogen, hydrogen, oxygen and carbon content. In order to be compatible with this technology, this work applies the support vectormachine (SVM) algorithm, which is particularly effective for statistical forecasting of small sample sets, to discriminate common household products and explosives according to the nitrogen, hydrogen, oxygen and carbon contents of the sample, Comparison of SVM, Fisher and Artificial Neural Networks. The results show that the SVM algorithm has the least false positives, and none of the explosives reported. Based on this, a prototype of the explosive discriminating system software was established, and the simulation test result was good in the laboratory.