Super-resolution Tomography Of Materials Based On Terahertz Time-domain Spectroscopy

In recent years,terahertz(THz)waves have attracted more and more attention due to its unique advantages.The terahertz nondestructive testing technology represented the terahertz time-domain spectroscopy(THz-TDS)and its imaging technique is expected to develop a new complementary method of nondestructive testing.In this paper,the terahertz nondestructive testing method is compared with the standard methods(X-ray industrial computed tomography and ultrasonic C-scan imaging)used in testing composite materials.The results show that the terahertz nondestructive testing technology has a higher image contrast but a lower spatial resolution in the transverse and axial directions.The ability to detect the acreage and thickness of defects in materials needs to be improved,and the method of automatic identification and quantification of defects also needs to be supplemented.Therefore,three high-resolution tomography methods based on THz-TDS technology are proposed and applied to the detection and quantification of defects in composite products,successfully detected and characterized the internal unknown defects,and verified the accuracy and practicability of the methods.The specific content is summarized as follows:(1)An adaptive window-segmentation quantization method based on THz-TDS technique improves the ability of acreage detection of defects.The peak detection method is used to automatically locate the reflected pulse related to the defect information,and the window function is used to extract the time domain pulse,and then the peak-to-peak parameter is used to map,which can achieve the maximum contrast imaging between the defect and the background.Combined with image segmentation and morphological denoising method,target defects can be identified automatically,and then the acreage can be detected automatically by using pixel number technology.For the detection of standard sample with defects,the proposed method can detect defects that cannot be detected by traditional methods,and the average detection error is 14,3%after automatic quantitative measurement of defect area.(2)Hilbert transform(HT)and Power spectrum estimation(PSE)based on THz-TDS technique is combined to improve the ability of thickness detection of defects.HT combined with PSE method is applied to the time domain segment where the defect reflection pulse is located.The pulse width can be reduced without changing the position of peak delay,the side lobe peak can be suppressed,and the adjacent reflection pulse can be distinguished successfully.Processing the time-delay parameters of peak-to-peak position,the thickness information is mapped to realize three dimensional information imaging.When applied to material detection,the method can effectively distinguish the upper and lower surface reflected pulses of 50?m thick delamination defects,and the average thickness detection error is reduced to 1,3%,(3)The fusion of terahertz and X-ray nondestructive testing technology further improves the ability of acreage detection of defects.In this paper,the high resolution advantage of X-ray nondestructive testing technology is introduced into terahertz nondestructive testing field for the first time to further improve the ability of defect detection and characterization.A fusion technique for terahertz nondestructive testing method and X-ray terahertz nondestructive testing method is developed.It is an image fusion algorithm based on significant region analysis(SRA)and wavelet based multi-scale transform(W-MST).The fusion algorithm optimizes the performance of the fusion algorithm by selecting fusion rules for constructing saliency maps,selecting decomposition levels of saliency maps and the wavelet basis,and designing fusion rules for high frequency and low frequency sub-bands.The objective evaluation indexes are used to quantitatively compare different combination algorithms,determining the optimization direction and choosing the most optimal algorithm.Finally,the most optimal fusion algorithm is applied to material detection,and the improvement effect of fusion method on defect detection is analyzed visually.After image segmentation and quantization,the results show that the fusion method can further reduce the average error of defect acreage detection to 5.0%,(4)The proposed THz imaging methods are applied to detect the glass fiber pipe products and the foam core material products,realizing the imaging of materials inside the unknown defects.The defect imaging results are highly consistent with the results of X-ray computer tomography imaging with high spacial resolution,proving the accuracy and practicability of the proposed methods in nondestructive testing.In this paper,the high resolution tomography methods based on THz-TDS technology can improve the lateral and axial resolution,and can effectively expand the detection limit of defects when they are applied to material detection.Further,after image segmentation processing,the methods can effectively realize automatic quantitative detection of defects and accurately measure the area and thickness of defects,providing technical support for nondestructive testing of composite materials.