| Terahertz radiation falls in between infrared radiation and microwave radiation in the electromagnetic spectrum,and it shares some properties with each of these.Terahertz radiation can penetrate fabrics and plastics,so it can be used in surveillance,such as security screening,to uncover concealed weapons on a person,remotely.This is of particular interest because many materials of interest have unique spectral "fingerprints" in the terahertz range.This offers the possibility to combine spectral identification with imaging.This thesis study explosives detection based on THz technology and implement the signal process algorithm.First,Chap 2 introduces how THz waves are generated and detected.The most common way that pulsed systems work is by splitting a beam from a femtosecond(fs)laser into two beams:the probe and the pump beams.Once appropriate THz wave sources and detectors become available,researchers will be interested in developing application technologies,in order to utilize the THz waves.We briefly introduce a few spectroscopic and imaging techniques.Second,Component analysis of spatial and spectral patterns are introduced to detect explosives signal which generated by manual.And improved by feeding reference signal to reduce computing burden.Further discuss the influence of threshold when using correlation filter to detect weak explosives signal.Then,because Component analysis of spatial and spectral patterns is unstable when explosives signal are weak.So ICA are introduced to extract explosives signal in mixed signal.A variant are introduced base on the cICA which separates only the target component.It shows that the presence of explosives can be accurately determined within the filed.Last,the advantage and disadvantage of CASP and ICA are discussed and compared.An improved method which utilizes both algorithms are proposed.The results obtained are promising and the approach has the capability to detect the explosives hidden under some background. |
PDF Full Text Request