Research On Overlapped Material Recognition Of Universal Security Inspection Equipment And R-l Object Classification

Dangerous goods inspection device is made for eliminating the dangerous goods caused by the insecurity of the social security needs of the developed, it has been widely used in airports, railways, ferry, road traffic and other important places, and also be available for postal services, conferences and special departments, and other security checks.It is the use of X-ray imaging technology went objects, access to the original digital images array data graphics and video processing and experimental building material classification by classification regional classification threshold, defineding organic, inorganic, compounds further distinguishing between dangerous goods. Because the equipment is technology intensive, advanced imaging technology, it can be effective for the identification of dangerous goods and enhance the quality of Dangerous Goods curse. Its success will surely enhance road transport safety, the protection of lives and property of the people and social stability, anti-terrorism, happenings, etc. will play an important role.Because the traditional X-ray inspection can only detect the higher Zeff materials, for example metals, it is unable to detect the lower Zeff materials, especially the explosives. So the dissertation is for this motivation.For detecting the explosives, we must enhance the ability of classifying materials. So the author made a modified system of the DEX10080B security inspection which is made by Company Di Tai in Northeastern University. By adding a forward scattering detection, a backward scattering detection and some hardware electrocircuits which are compatible to the detection, it is formed to a double energy transmission and scattering detection system. Based on the system, the author designed an image system and overlapped models of transmission, forward scattering and backward scattering to gain the true gray levels of the objects, and so we can obtain the aim of recognizing overlapped materials. Beside these, the author used Numerical Optimization Method to gain value R which is relative to Zeff, and LS method to gain value L which is relative to material's density, and through the experiments of explosives, organics, inorganics gain value R and L. The author designed the discriminant function and decision boundary of explosives and organics by using Bayes decision theory based on the minimum-error-rate classification and build R-L characteristic special curve. This method greatly enhances the ability of classifying materials of X-ray double energy and scattering.