Research On Terahertz Time Domain Spectral Imaging Algorithm Based On Compressed Sensing

Terahertz Time-Domain Spectroscopy(THz-TDS)is an emerging spectroscopy technology that uses terahertz pulsed electromagnetic waves to reflect material properties.It is widely applied in the fields of biomedicine,non-destructive testing and so on.THz-TDS imaging is becoming a complementary technology to ultrasound imaging,X-ray imaging and nuclear magnetic resonance imaging,but THz imaging currently has the problem of large data acquisition and huge time costs,which severely limits the research process of THz-TDS technology application.At present,the research on terahertz fast imaging technology mainly focuses on the improvement of hardware,which not only increases the cost of imaging,but also increases the complexity of the system.On the other hand,only a few researches have been carried out which aim at imaging algorithms based on THz characteristics,and there is still a lack of effective THz fast imaging algorithms.In this thesis,the under-sampling method of signal is employed to study the terahertz spectrum and imaging algorithm based on compressed sensing,and the effectiveness of the algorithm is verified by programming.The main research work and innovations are as follows:(1)The research on one-dimensional compressed sensing algorithm for THz-TDS signal is carried out.Combined with the characteristics of THz spectral signals,the improved evaluation method of reconstruction algorithm based on calculation of material property parameters is proposed.The spectral data of the water-containing cardboard was obtained by THz-TDS,and the sparseness of the THz spectral signal was verified in the thesis.The THz spectral signal was reconstructed by the orthogonal matching pursuit algorithm.At the same time,the evaluation basis of the algorithm was corrected based on the physical property.The result shows that:(a)the improved evaluation method can more accurately describe the effectiveness of the reconstruction algorithm,ensuring the integrity of material parameter information;(b)using the compressed sensing algorithm,initial signals can be almost completely reconstructed using only 27% of the data.The research gives the formula of the minimum number of measurements,which provides a theoretical basis for quickly determining the sampling rate.(2)Based on the transmissive THz-TDS system,a two-dimensional imaging analysis of terahertz is carried out.The principle of THz-TDS imaging and its unique imaging modes are studied.These imaging experiments were carried out on the samples of insulating dielectric materials and biological materials which are very different in physic properties,so that the effectiveness and universality of THz-TDS systems can be verified,and the function of different imaging modes were illustrated by examples.The experimental result shows that:(a)THz imaging can effectively detect air gap defects in XLPE,and time slice mode can effectively improve imaging contrast;(b)THz imaging technology can identify skin burns,using flight time imaging mode can effectively filter out background noise and highlight the boundaries of the burned area.(3)Research on two-dimensional compressed sensing imaging algorithm for THz spectral imaging.The TVAL3 algorithm based on fast Hadamard transform is proposed.The comparison experiment shows that TVAL3 reconstruction algorithm has obvious advantages in the processing of two-dimensional THz image than OMP algorithm.When the THz image data is processed by the TVAL3 algorithm,the original image can be reconstructed with only a small amount of data of nearly 20%,which greatly reduces the amount of data collected.Further research shows that combined with the inherent characteristics of THz spectroscopy,the appropriate THz imaging mode can significantly improve the processing effect of the compressed sensing imaging algorithm,and further reduce the signal acquisition ratio.