Terahertz Spectral Imaging Technology And Image Enhancement Technology

Terahertz wave has the characteristics of low energy,coherence,high penetration,transient,fingerprint spectrum and so on.At the same time,CT imaging can provide the specific information of the internal structure.CT based on THz-TDS and spectral tomography technology have a good development potential and application prospect in the material sample detection.Terahertz time-domain spectral imaging technology can not only distinguish the shape of the sample,but also distinguish the components through the pixel information of different positions.In this paper,based on the theory of terahertz spectroscopy,terahertz time domain spectroscopy(THz-TDS)system is used to obtain the spectral data of samples.Two key points in sample detection are studied.First of all,using the spectral information of the sample,we use terahertz CT imaging technology to obtain the internal structure of the sample.At the same time,we use the point spread function constructed by the physical characteristics of terahertz beam to deconvolute the terahertz image to improve the resolution.Secondly,based on the work of the first part,we make full use of the spectral information of each pixel of the sample image to analyze the material information of the sample,so as to obtain the material information of the sample.The main research contents and innovations are as follows:Frist,A method of terahertz sample recognition based on neural network is researched.Terahertz wave is very sensitive to the small changes of material structure and environment.Terahertz time-domain spectroscopy has great potential in the study of internal structure and molecular interaction of materials.It can give the unique fingerprint spectrum of objects,and the terahertz spectrum of objects with very similar structure is also very different,so it can be used for object classification.The method mentioned in this paper classifies and identifies the detected objects by extracting various characteristic parameters such as refractive index,absorption index,dielectric constant and material factor.By calculating its absorption spectrum,refractive index,flight time and other characteristics,BP,SVM and LVQ are used to detect drugs with multiple characteristics.The accuracy can reach more than 95%,which can be used to detect and identify drugs.Then,A fast convergence method of terahertz CT is researched.Terahertz imaging can obtain multi-layer information of objects,which is an important direction of terahertz technology development.MLEM algorithm based on maximum likelihood is a classical imaging method,which has better imaging quality.In principle,the imaging result is better than filter back projection algorithm.The order subsets OSEM algorithm improves the convergence speed greatly,so OSEM algorithm and its various deformation algorithms are widely used.However,the inherent characteristics of terahertz increase the complexity of the algorithm,and the use of OSEM algorithm for terahertz imaging still needs many iterations,reducing the imaging efficiency.Therefore,the iterative process of terahertz imaging is time-consuming,and the convergence speed needs to be further improved.In this paper,by combining the ordered subsets and the quadratic step convergence imaging method,the convergence speed of the algorithm is further improved while the imaging quality is guaranteed,which improve the imaging efficiency.Finally,A terahertz 3D image reconstruction method is researched.In this paper,THz timedomain spectral projection data of samples are measured by THz time-domain spectrometer,and three-dimensional reconstruction of samples is carried out.Combined with the advantages of THz wave safety and high resolution,a new idea is provided for the nondestructive testing of samples.